THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society for Biochemistry and Molecular Biology, Inc.

Val. 262, No. 24, Issue of August 25, pp. 11651-11656,1987 Printed in U.S.A.

Primary Structure Requirements for the Binding of Human High Molecular Weight Kininogen to Plasma Prekallikrein and Factor XI*

(Received for publication, March 23, 1987)

Jonathan F. TaitSQ and Kazuo FujikawaS From the Departments of $Biochemistry and §Laboratory Medicine, University of Washington, Seattle, Washington 98195

We recently identified residues 185-224 of the light chain of human high molecular weight kininogen (HMWK) as the binding site for plasma prekallikrein (Tait, J. F., and Fujikawa, K. (1986) J. Biol. Chen. 261, 15396-15401). In the present study, we have further defined the primary structure requirements for binding of HMWK to factor XI and prekallikrein. In a competitive fluorescence polarization binding as- say, a 31-residue synthetic peptide (residues 194-224 of the HMWK light chain) bound to prekallikrein with a K d of 20 f 6 nM, indistinguishable from the previ- ously determined value of 18 f 5 nM for the light chain. We also prepared three shorter synthetic peptides cor- responding to different portions of the 31-residue pep- tide (residues 205-224,212-224, and 194-211), but these peptides bound to prekallikrein more than 100- fold more weakly. Factor XI also bound to the same region of the HMWK light chain, but at least 58 resi- dues (185-242) were required for optimal binding ( K d = 69 f 4 nM for the light chain; Kd = 130 f 50 n M for residues 185-242). The four synthetic peptides inhib- ited kaolin-activated clotting of blood plasma with po- tencies paralleling their affinities for prekallikrein and factor XI. Peptide 194-224 can also be used for rapid affinity purification of prekallikrein and factor XI from plasma.

Four proteins initiate blood coagulation in the presence of anionic surfaces: factor XII, prekallikrein, factor XI, and high molecular weight kininogen (HMWK)’ (1,2). Factor XI1 and prekallikrein undergo reciprocal proteolytic activation; factor XIIa then activates factor XI, which activates the remainder of the clotting cascade (3, 4). HMWK greatly accelerates these reactions by functioning as a nonenzymatic cofactor (5, 6). Kallikrein and factor XIIa can also activate elements of the fibrinolytic system (7) and the inflammatory response (8).

HMWK participates in contact activation reactions through specific interactions with prekallikrein and factor XI. Both of these factors circulate in plasma as complexes with HMWK (9, 10). HMWK promotes the binding of prekalli- krein and factor XI to kaolin (11). In direct binding studies, HMWK binds tightly to prekallikrein (12-14) and factor XI (12, 15, 16), which compete for the same site on HMWK (12,

* This work was supported in part by Research Grant HL 16919 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The abbreviations used are: HMWK, high molecular weight ki- ninogen; factor XIIa, activated factor XII; FITC, fluorescein isothi- ocyanate; HEPPS, N-2-hydroxyethylpiperazine-N’-3-propanesul- fonic acid; light chain-2, residues 49-255 of the HMWK light chain; Polybrene, hexadimethrine bromide.

15). This binding region has been localized to the HMWK light chain (12, 13, 15, 17, la), somewhere in the COOH terminal 123 residues of the bovine light chain (19).

Recently, the complete primary structures of the four hu- man contact factors have been determined (factor XI1 (20, 21), prekallikrein (22), factor XI (23), and HMWK (24-26)). It is therefore of interest to study the relationship between the observed primary structural features and the functions of these proteins. We have shown that the binding site for prekallikrein in HMWK is fully contained in a 40-residue tryptic fragment derived from the near carboxyl terminus (residues 185-224 of the light chain) (27). This fragment has a dissociation constant of 12 f 5 nM, equivalent to the value for the intact light chain (27).

In this study, we have used synthetic and naturally derived peptides to define the requirements for optimal binding of HMWK to plasma prekallikrein and factor XI. We show that a 31-residue synthetic peptide retains full binding activity for prekallikrein, while at least 58 residues are required for opti- mal binding to factor XI. This synthetic peptide inhibits kaolin-activated clotting of plasma and can also be used as an affinity ligand for the rapid purification of prekallikrein and factor XI from plasma.

EXPERIMENTAL PROCEDURES

Materials-Fresh frozen human plasma was kindly provided by American Red Cross Blood Services, Portland, OR. Other materials were from these sources: solvents and amino acid derivatives for peptide synthesis (Applied Biosystems); tresyl-activated Sepharose 4B (Pharmacia); human coagulation factor-deficient plasmas (George King Biomedical); bovine factor XI-deficient plasma (kindly provided by Dr. Kociba of Ohio State University); fluorescein isothiocyanate “isomer I” (FITC) (Molecular Probes); benzoyl-Pro-Phe-Arg-p-ni- troanilide (VegaBiochemicals); D-Phe-L-Phe-L-Arg-chloromethyl ke- tone (Behring Diagnostics); disposable acrylic fluorescence cuvettes (Sarstedt); Polybrene (Aldrich).

The following E% and M, values were used prekallikrein, 11.7 (17) and 82,000 (calculated from the known amino acid composition (22) and carbohydrate content (28)); factor XI (dimer), 13.4 (29) and 143,000 (calculated from the known amino acid composition (23) and carbohydrate content (29)).

Protein and Peptide Purification-HMWK was prepared from human plasma by a modification of the method of Kat0 et al. (30) as previously described (27). The HMWK light chain and its fragments were prepared as described (27); the fragments used in this study were light chain-2 (residues 49-255), CNB-111 (residues 185-242), T- 7 (residues 185-224), and CNB-I1 (mixture of residues 49-184 and

Synthetic peptides were kindly prepared by Dr. Patrick Chou (Howard Hughes Medical Institute, University of Washington) by solid phase synthesis on an Applied Biosystems 430A peptide syn- thesizer from phenylacetamidomethyl resins and t-butoxycarbonyl amino acids. For preparation of the peptide-IV affinity column (see below), the synthesized product was used without further purification. For binding studies, peptides were purified by reverse-phase high performance liquid chromatography on a C,-Cls column (Pharmacia Biotechnology Inc. PepRPC 0.5 X 5 cm) with a gradient between

243-255).

11651

11652 Binding of HM WK to Prekallikrein and Factor X I 0.1% trifluoroacetic acid and 0.08% trifluoroacetic acid in 80% ace- tonitrile. The amino acid compositions of the four synthetic peptides used in this study were consistent with the intended compositions within 510% (data not shown). Peptide-I11 was also sequenced by automated Edman degradation (Beckman model 890C) for 18 cycles and yielded the intended sequence for all 18 residues (not shown).

FITC-peptide-IV was prepared by incubating peptide-IV (200 p ~ ) with equimolar FITC for 24 h at 37 "C in 0.15 M sodium borate, pH 9.0. The labeled peptide was separated from unlabeled peptide and FITC by reverse-phase high performance liquid chromatography as described above. The estimated labeling ratio was 2.1 mol of fluores- cein per mol of peptide, determined from amino acid analysis and an

Concentrations of all peptide stock solutions were determined by amino acid analysis of 24-h hydrolysates.

Preparation of Plasma Prekallikrein and Factor XI by Affinity Chromatography with Peptide-ZV-Sephrose-Fresh frozen plasma

Tris-HC1, pH 7.5, 20 mM EDTA, 20 mM benzamidine-HCl, and 100 (1.5 liters) was thawed and diluted with an equal volume of 50 mM

mg/liter of Polybrene. After centrifugation, the diluted plasma was applied to a column (3.2 X 10 cm, 80 ml) of peptide-IV-Sepharose that had been equilibrated with 50 mM Tris-HC1 buffer, pH 7.5, containing 75 mM NaCl, 10 mM EDTA, 10 mM benzamidine-HC1, and 50 mg/liter of Polybrene. Following a wash with 0.5 liter of the equilibration buffer, adsorbed proteins were eluted first with 300 ml of 0.1 M citrate buffer, pH 5.0, containing 1 M NaCl, 10 mM benz- amidine, 10 mM EDTA, and 50 mg/liter of Polybrene. This buffer eluted most of the factor XI activity and part of the prekallikrein activity. The remaining prekallikrein was then eluted with 300 ml of 0.1 M citrate buffer, pH 3.5, containing 1 M NaCl, 0.2 M guanidine hydrochloride, 10 mM benzamidine, 10 mM EDTA, and 50 mg/liter of Polybrene. The eluate was collected in tubes containing 0.1 volume of 1 M Tris-HC1, pH 8.0. Both eluates were combined and dialyzed twice against 6 liters of 50 mM Tris-HC1, pH 7.5, containing 50 mM NaCl, 10 mM benzamidine, and 50 mg/liter of Polybrene. The dialy- sand was adjusted to a conductivity of 5 mmho and applied to a column (3.2 X 24 cm) of heparin-agarose equilibrated with the same buffer. After a 1-liter wash, proteins were eluted with a linear gradient composed of 1 liter each of 50 mM and 500 mM NaCl in the same buffer, and LO-ml fractions were collected. Every fifth fraction was assayed for prekallikrein (amidolytic activity) and factor XI (clotting activity) as described below.

The fractions containingprekallikrein were dialyzed against 4 liters of a buffer containing 50 mM sodium acetate, pH 5.2, and 50 mM NaC1, and then applied to a column (0.9 X 10 cm) of CM-Sephadex previously equilibrated with the dialysis buffer. The column was washed with 200 ml of 50 mM sodium acetate, pH 5.2, containing 150 mM NaCl, and then eluted with a gradient composed of 75 ml each of 150 mM and 350 mM NaCl in 50 mM acetate buffer, pH 5.2. The fractions containing prekallikrein activity were pooled, concentrated to approximately 10 ml using an Amicon ultrafiltration apparatus with a PM-10 membrane, and stored frozen at -80 "C. The yield of prekallikrein was approximately 4 mg from 1.5 liters of plasma, with a clotting activity of 45-65 units/mg.

The factor XI fraction from the heparin-agarose column was dialyzed against 20 mM phosphate buffer, pH 7.2, containing 70 mM NaCl and 10 mM benzamidine and then applied to a column (2 X 7 cm) of CM-Sephadex equilibrated with the dialysis buffer. Following a 200-ml wash with the same buffer, factor XI was eluted with a gradient composed of 75 ml each of 70 mM and 600 mM NaCl in 20 mM phosphate buffer, pH 7.2, containing 10 mM benzamidine. Frac- tions containing factor XI activity were pooled, dialyzed against 50 mM sodium acetate buffer, pH 5.5, concentrated to approximately 5 ml, and stored frozen at -80 "C. The yield of factor XI was approxi- mately 1 mg from 1.5 liters of plasma, with a clotting activity of 160- 220 units/mg.

Peptide-IV-Sepharose was prepared by coupling 200 mg of peptide- IV to 8 g of tresyl-activated Sepharose according to the manufactur- er's instructions. Heparin-agarose was prepared as described earlier (32). 8-Factor XIIa was prepared as described (20).

Clotting Assays-For assay of prekallikrein and factor XI, test samples were diluted 10-fold with 50 mM Tris-HC1 buffer, pH 7.4, containing 150 mM NaCl and 1 mg/ml bovine serum albumin. Twenty p1 of test sample was incubated for 10 min at 37 "C with 20 pl of kaolin (5 mg/ml in normal saline) and 20 p1 of factor-deficient plasma (human for prekallikrein; human or bovine for factor XI). The clotting time was determined visually following addition of 40 pl of an equal mixture of 0.033 M CaCL and cephalin (the contents of one

6494 Of 78,000 M" cm" (31).

vial of Sigma rabbit brain cephalin were suspended in 100 ml of saline and stored in aliquots at -20 "C). One unit is defined as the amount of clotting activity of prekallikrein or factor XI present in 1 ml of pooled normal human plasma.

To study the effect of synthetic peptides on clotting of blood plasma, samples were prepared in 60 pl with the indicated final concentration of synthetic peptide and the following final concentra- tions of other reagents: HMWK-deficient plasma (1:3 dilution), HMWK (10 nM), kaolin (1.7 mg/ml), sodium MOPS (150 mM, pH 7.4). After a 2-min incubation at 37 "C, the clotting time was deter- mined after addition of 40 p1 of rabbit brain cephalin in 16 mM CaC12, 75 mM sodium MOPS, pH 7.4, 75 mM NaCl. Under these conditions, the control clotting time was 150 s.

Assay of Preknllikrein Amidase Activity-One hundred pl of 0.5 M Tris-HC1, pH 8.0, was added to test samples (20-100 pl), and prekal- likrein was activated to kallikrein by incubation for 10 min at 37 ' with 2 pg of @-factor XIIa. One hundred pl of 0.1 mM benzoyl-Pro- Phe-Arg-p-nitroanilide was added, and the reactions were terminated at appropriate times by the addition of 10% acetic acid to a total of 1 ml. The amidase activity of kallikrein was calculated from the absorbance change at 405 nm.

Fluorescence Polarization Binding Assay-Binding measurements were performed at 34 "C in a buffer consisting of 0.1 M HEPPS, 0.05 M NaC1,O.l mM EDTA, 0.1% polyethylene glycol 6000, pH 8.0, ionic strength, 0.10. (Polyethylene glycol was used to minimize adsorption of proteins to the cuvette (17).) For experiments with the HMWK light chain or light chain-2, the buffer also contained 10 pM D-Phe- L-Phe-L-Arg-chloromethyl ketone and 0.1 mM diisopropyl fluoro- phosphate. For experiments with prekallikrein, disposable acrylic cuvettes were used; for factor XI, a glass microcell was used. Polari- zation measurements were made in T format on an SLM-8000C fluorometer (SLM/Aminco, Urbana, IL) with excitation at 490 nm (16-nm slit) and emission monitored through Schott (Duryea, PA) KV-550 long-pass filters (50% transmission at 550 nm). Excitation intensity was varied as necessary with neutral density filters. Solvent blanks were automatically subtracted to yield net fluorescence inten- sity and polarization. Polarization readings were corrected for optical and electronic bias with horizontally polarized excitation light ac- cording to standard methods (33), and fluorescence intensities were corrected for dilution during titrations. Polarization values were used to calculate a, the bound fraction of FITC-peptide-IV (33):

P - P o (P1m - P)R + P - P o

a = (1)

where P is the observed polarization, Po is the polarization of ligand in the free state, Plm is the polarization of ligand in the bound state, and R is the ratio of relative fluorescence intensities of the bound and free forms of the ligand. Po was measured directly; Pt00 and R were determined by fitting Equation 4 (below) to the observed polar- ization and fluorescence intensity changes in the direct titrations. The following average values (fSD) were obtained PO, 0.0505 f 0.0019 ( n = 69); for prekallikrein, Plw = 0.1797 f 0.0038 and R = 1.17 f 0.02 (n = 7); for factor XI, Plm = 0.1763 f 0.0116 and R = 1.32 f 0.08 (n = 8).

Binding constants were determined either by direct titration of fluorescent ligand with prekallikrein or factor XI, or by competition assay with unlabeled ligand. In the direct titrations, increasing amounts of prekallikrein or factor XI were added sequentially to a fixed amount of FITC-peptide-IV (1-18 nM for prekallikrein, 5-2000 nM for factor XI), and the fluorescence intensity and polarization were recorded after each addition. In competition assays, increasing amounts of unlabeled competitor were added to fixed amounts of FITC-peptide-IV and prekallikrein or factor XI, and polarization was recorded after each addition.

Data Analysis-Binding constants were determined from the fol- lowing model (essentially that of Bock and Shore (17)):

where A represents FITC-peptide-IV, B represents unlabeled com- petitor, and C represents either prekallikrein or factor XI (as mon- omer). This model assumes that A and B compete for the same site on C with the same stoichiometry m. Previous work has demonstrated a 1:l stoichiometry for the prekallikrein-HMWK interaction (13,14,

Binding of HMWK to Prekullikrein and Factor XI 11653

17.18) and the factor XI monomer-HMWK interaction (16), but the stoichiometry factor m was introduced to allow for the presence of inactive protein or systematic errors in concentrations (see “Re- sults”).

In direct titrations, the dissociation constant F d was determined by fitting a, the fraction of A bound to C, to the quadratic equation describing Reaction 2 above:

a = (1/(2At))[K’d + At + mCt (4)

- ((K’d + At + mc,)’ - 4AtmC,)”*]

where A, is the total concentration of A and C, the total concentration of C. Fitting was performed by unweighted nonlinear least squares regression with the RS/1 software package (BBN Software Products, Cambridge, MA).

In competition experiments, the dissociation constant Kd was determined according to a modification of Equation 18 of Dandliker et al. (34) as previously described (27). In this approach, a (and hence the bound/free ratio for A) is determined from Equation 1; Kd is then calculated from the bound/free ratio, the average value of K l d , and the known total concentrations of A, B, and C. The concentration of C was adjusted by the stoichiometry factor m determined by direct titration ofA with C. In each experiment, an average Kd was calculated from 4-10 individual values corresponding to 15-85% displacement of the labeled ligand.

RESULTS

Binding of Synthetic Peptides to Prekallikrein-Four pep- tides were synthesized (Fig. 1) to correspond to different portions of the 40-residue tryptic peptide T-7, which was previously found to retain full binding activity for prekalli- krein (27). We focused on residues 194-224 of T-7 because this subregion has the highest homology with bovine HMWK (Fig. 1). Preliminary experiments showed that peptide-IV had the highest affinity for prekallikrein of the four synthetic peptides when assayed according to our previous methods (27). This peptide was therefore labeled with FITC and used for direct binding studies; FITC-peptide-IV was preferable to FITC-light chain-2 because it was not susceptible to proteol- ysis, did not adsorb to the cuvette, and gave a much larger polarization change on binding. Titration of FITC-peptide- IV with prekallikrein (Fig. 2) caused a large increase in fluorescence polarization (open circles) and a slight increase

Heavv chain

(17%) in fluorescence intensity (not shown). After correction for the difference in fluorescence intensity of free and bound forms of FITC-peptide-IV (closed circles), the data could be well fit by a model of a single class of noninteracting binding sites (solid line through solid circles). The following parame- ters were obtained from a series of titrations of FITC-peptide- IV with a single preparation of prekallikrein: K l d , 9.1 _+ 0.9 nM ( n = 7); m, 0.75 -+ 0.10 ( n = 3) (mol of FITC-peptide-IV/ mol of prekallikrein). This K d is indistinguishable from val- ues we have previously obtained for FITC-T-7 (7 nM) or FITC-light chain-2 (12 nM) (27). The nonintegral stoichi- ometry is most likely due to the presence of some denatured prekallikrein.

The affinities of the synthetic peptides for prekallikrein were determined in a competition assay (Fig. 3). In this assay, increasing amounts of unlabeled competitor were added to fixed amounts of FITC-peptide-IV and prekallikrein. Peptide- IV (open circles) had the same molar potency as light chain- 2 (closed circles), confirming the results obtained from the direct titration of FITC-peptide-IV. However, the three shorter peptides (crosses, squares, triangles) showed more than 100-fold lower potency than either light chain-2 or peptide-IV.

Binding of Natural and Synthetic Peptides to Factor XI- Factor XI also bound to FITC-peptide-IV (Fig. 2) with ap- proximately the same polarization increase on binding (open squares) as prekallikrein (open circles). However, the fluores- cence intensity change was larger (a 32% increase, uersus 17% for prekallikrein (not shown)), indicating differences in the environment of the fluorophore in the two types of protein- peptide complexes. After correction for changes in fluores- cence intensity (solid squares), the data for factor XI could also be well fit by a model of homogeneous independent sites (solid line). We obtained an average K f d of 750 f 310 nM from 10 titrations of FITC-peptide-IV with four preparations of factor XI. A stoichiometry m of 0.82 & 0.11 (mol of FITC- peptide-IV/mol of factor XI monomer) was obtained in three titrations, again suggesting the presence of some denatured protein.

Bradykinin Light chain HMWK I i

1 371/1 185 242 255 c”-------1 CNB-I I I

CNB-I11 I l / - l

/M T-7 I

Peptide-IV (194-224) I I

Peptide-I11 (205-224) c I

Peptide-I1 (212-224) I I

Peptide-I (194-211) I I

FIG. 1. Location of natural and synthetic peptides within the primary structure of HMWK. The schematic diagram at the top of the figure shows the principal primary structural features of human HMWK and the location of fragment CNB-111. Next is shown a portion of the amino acid sequence of the human light chain (24) (upper) with the corresponding region of the bovine light chain (43,44) (lower). The names and locations of the natural and synthetic peptides are given below. Residue numbers refer to the human HMWK heavy and light chains (24,26).

11654 Binding of HMWK to PI

[Protein]totol ( n M 1 FIG. 2. Titration of FITC-peptide-IV with prekallikrein

(circles) and factor XI (equares). FITC-peptide-IV (4.6 nM for prekallikrein, 10.5 nM for factor XI) was titrated with prekallikrein or factor XI as described under "Experimental Procedures." The open symbols and dotted lines are changes in net fluorescence polarization relative to unbound FITC-peptide-IV; closed symbols represent the percentage of FITC-peptide-IV bound to protein (calculated accord- ing to Equation 1); solid lines represent the fit of Equation 4 to the percentage bound. The concentration of factor XI is given as the monomer. For these data, the fitted K d values are 8.9 nM for prekal- likrein and 430 nM for factor XI.

t 1 I I O 0 10000

(nM)

FIG. 3. Binding of synthetic peptides to prekallikrein. Cu- vettes were prepared with 14.6 nM prekallikrein and 4.6 nM FITC- peptide-IV. Competitors were then added repetitively to give the indicated final concentration, and polarization readings were made after each addition. The amount of bound FITC-peptide-IV is given as a percentage of the amount bound in the absence of competitor. Symbols: filkd circle, light chain-2; open circle, peptide-IV; x, peptide- 111; open square, peptide-& open triangle, peptide-I.

To investigate the primary structure requirements for the binding of HMWK to factor XI, we performed competition assays with cyanogen bromide and tryptic fragments of light chain-2 (prepared as previously described (27)) (Fig. 4). For the natural fragments, the intact light chain and light chain- 2 had similar affinities for factor XI (left diamonds uerszu circles). Among the cyanogen bromide fragments, CNB-I11 (squares) (residues 185-242) bound nearly as tightly as light chain-2, while CNB-I1 (inverted triangles) (consisting of res- idues 49-184 and 243-255) was inactive. Tryptic fragment T- 7 (triangles) (residues 185-224) was also active, but less so than CNB-111. For the synthetic peptides (Fig. 5), peptide-IV (open circles) was 30-fold less potent than light chain-2 (closed circles), while the shorter peptides (crosses, squares, triangles) were some 2000-fold less potent.

Comparison of K d Values for Prekallikrein and Factor XI- Table I summarizes mean & values for the binding of the

*ekallikrein and Factor XI T ' . 1 . 100 . T .

L I 10 1 0 0 1000

[Competitor]+,t,( (nM) FIG. 4. Binding of HMWK light chain and its fragments to

factor XI. Competition assays were performed as described in the legend to Fig. 3, with 280 nM (as monomer) factor XI and 6.8 nM FITC-peptide-IV. Symbols: filled left trhgle, light chain; filled circle, light chain-2; filled square, CNB-111; filled triangle, T-7; filkd inverted triangle,CNB-11.

t L \. 1 10 1000 100000

[Cornpetitor]+,t,l (nM) FIG. 5. Binding of synthetic peptides to factor XI. Competi-

tion assays were performed as described in the legend to Fig. 3, with 280 nM (as monomer) factor XI and 6.8 nM FITC-peptide-IV. Symbols are as in Fig. 3. The data for light chain-2 are replotted for comparison from Fig. 4.

HMWK light chain and its fragments to prekallikrein and factor XI. As previously noted, the 31-residue peptide-IV was fully as active as light chain-2 for binding to prekallikrein, while at least 58 residues were required for optimal binding to factor XI. The binding of factor XI to the light chain and its fragments was also somewhat weaker than corresponding values for binding to prekallikrein. FITC-peptide-IV bound to prekallikrein and factor XI about %fold more strongly than did unlabeled peptide-IV, probably reflecting the electrostatic effect of additional negative charges due to the fluorescein moiety.

Effect of Synthetic Peptides on Clotting of Blood Plusma- We tested whether the synthetic peptides can act as antago- nists of contact activation reactions by adding each peptide to a HMWK-deficient plasma supplemented with a limiting amount of HMWK (10 nM). Under these conditions, clotting activity would be detected as a shortening of the clotting time, while inhibition would be seen as a prolongation. All peptides prolonged the kaolin-activated clotting time in a concentra- tion-dependent manner (Fig. 6). The relative potency of the peptides also paralleled their potency in binding to prekalli- krein and factor XI, with peptide-IV being the most potent by far (doubling of control clotting time at 4 p M peptide-IV). Similar results were obtained for peptide-IV in normal plasma

Binding of HMWK to Prekallikrein and Factor X I 11655

TABLE I Affinities of natural and synthetic peptides for prekallikrein and factor X I

~-

Fragment Location Length (residues)

K d : prekallikrein factor XI

K d :

nM nM

Light chain 1-255 255 18 f 5‘ Light chain-2

69 f 4 24 f 5b 63 f 13 27 f 5b 130 f 50

49-255 207 CNB-111 185-242 58 T-7 185-224 40 12 f 5b 520 f 190‘

194-224 31 20 f 6‘ 1,840 f 650d Peptide-IV Peptide-I11 205-224 20 2,700 f 1,000 126,000 f 16,000 Peptide-I1 212-224 13 6,000 2 3,200 150,000 f 80,000 Peptide-I 194-211 18 8,500 f 2,100 CNB-I1

410,000 f 120,000 49-184 136 >>1,00Ob >>10,000

243-255 13 K d values are given as mean f SD for two or more separate competition experiments, each with 4-10 individual

determinations, performed as described under “Experimental Procedures.” Data from Figs. 3-5 were included in these calculations.

From Ref. 27. ‘ n = 3 . ’ n = 4 .

[Pe~tide]+~t,,l (I”

FIG. 6. Effect of synthetic peptides on kaolin-activated clot- ting of blood plasma. Clotting assays were performed as described under “Experimental Procedures” with HMWK-deficient plasma supplemented with 10 nM purified HMWK. The control clotting time was 150 s. Symbols are as in Fig. 3.

(not shown), although approximately 20-fold higher concen- trations of peptide were needed to achieve the same degree of inhibition.

Affinity Chromatography with Peptide-IV-HMWK-Seph- arose has been previously used in the purification of prekal- likrein (35) and factor XI (15). We investigated the usefulness of synthetic peptides for this procedure by preparing an affinity column of peptide-IV-Sepharose. Human plasma was applied directly to the column; after washing, proteins were eluted from the column as described under “Experimental Procedures.” The eluate showed only three major bands on SDS-polyacrylamide gel electrophoresis (Fig. 7, lane 5): IgG, factor XI dimer, and prekallikrein (in order from the top). Prekallikrein and factor XI co-migrated under reducing con- ditions ( l a n e 6). Prekallikrein and factor XI were separated by chromatography on heparin-agarose; each was then puri- fied by CM-Sephadex chromatography. The final products were pure prekallikrein and factor XI (Fig. 7, lanes 1-4), primarily or exclusively in single-chain form.

DISCUSSION

This study and our previous work (27) show that factor XI and prekallikrein share a common binding site in HMWK that is located in a short region near the carboxyl terminus of HMWK. The 31-residue peptide-IV contains the most

”94,000

“68,000

“43 000

--30,000

--14,400

FIG. 7. Affinity chromatography of plasma prekallikrein and factor XI on peptide-IV-Sepharose. Prekallikrein (lanes 1 and 2) and factor XI (lanes 3 and 4 ) purified on the affinity colnmn as described under “Experimental Procedures“ were subjected to SDS-polyacrylamide gel electrophoresis (45). Lanes 5 and 6 show the composition of the eluate from the peptide-IV-Sepharose column following application of plasma. The column was washed and eluted as described under “Experimental Procedures.” Lane 7 contains stan- dards with the indicated molecular weights. Odd-numbered lanes, no reducing agent; even-numbered lanes, samples reduced with 5% 2- mercaptoethanol.

important determinants for factor XI binding and all the determinants for prekallikrein binding (Fig. 1 and Table I). These results have been confirmed by two additional tech- niques: affinity chromatography with peptide-IV, which spe- cifically binds prekallikrein and factor XI in whole plasma; and plasma clotting assays, which show that peptides I-IV cause dose-dependent inhibition of clotting with potencies paralleling their affinities for the purified proteins in uitro.

We previously showed that the prekallikrein binding site in HMWK is located in the 40-residue tryptic fragment T-7 (residues 185-224 of the HMWK light chain) (27). It is now apparent that the 9 NH2-terminal residues of T-7 make no significant contribution to this binding site, since peptide-IV

11656 Binding of HMWK to Pretzallikrein and Factor X I

and T-7 bind to prekallikrein with equivalent potency (Table I). Results obtained with synthetic peptides also show that the presence of carbohydrate chains on Ser-188 and Thr-204 is not necessary for high affinity binding to prekallikrein. However, further subdivision of this active region led to dramatic decreases in binding activity (Table I). This active region shows very high homology with the corresponding region of the bovine protein (Fig. 1): 81% of residues 194-224 are identical between the two proteins, compared with 68% for the entire human and bovine light chains (24). As we have previously noted (27), this region is rich in acidic residues (7 Asp) and poor in basic residues (1 Lys).

Factor XI and plasma prekallikrein are highly homologous proteins, with 58% identity overall (22, 23), and previous studies have shown that factor XI, like prekallikrein, binds tightly to HMWK (12, 15, 16, 36) or the HMWK light chain (12,15,36). However, the detailed structural requirements for binding are somewhat different. Optimal binding of prekalli- krein to HMWK requires at most a 31-residue sequence (peptide-IV), while optimal binding to factor XI requires at least a 58-residue region (CNB-111). This region encompasses additional residues both NH2-terminal and COOH-terminal to the prekallikrein binding site. However, these additional 27 residues increase the strength of binding only l4-fold, indicating that the 31-residue region contains the most im- portant determinants for binding to factor XI.

Synthetic peptides which disrupt the complexes of HMWK with prekallikrein and factor XI offer new tools for investi- gation of the clinical and physiological roles of the contact system. Activation of the contact factors may be important in the pathogenesis of several disorders (37). Bradykinin, which is cleaved from HMWK by kallikrein, is an important inflammatory mediator (38). Levels of prekallikrein and HMWK fall during attacks of angioedema in the disease hereditary angioedema (39). HMWK and bradykinin are re- leased into the nasal fluid when allergic, but not nonallergic, subjects are challenged with allergens (40). Kallikrein causes neutrophils to aggregate (41) and to release elastase (42). We have shown that synthetic peptide-IV acts as an in vitro anticoagulant and can also bind directly to prekallikrein and factor XI in whole plasma. Because synthetic peptides can be easily prepared in large quantities, this may offer a new approach to the development of anticoagulant or anti-inflam- matory substances.

Acknowledgments-We would like to thank Dr. Earl W. Davie for his continuing support and encouragement of this work. We thank Dr. Patrick Chou for preparing the synthetic peptides, Don Gibson and Lee Hendrickson for excellent technical assistance, Brad MC- Mullen for sequence analysis, and Roger Wade for assistance with amino acid analysis.

REFERENCES 1. Kisiel, W., and Fujikawa, K. (1983) Behring Znst. Mitt. 73, 29-

2. Griffin. J. H.. and Cochrane, C. G. (1979) Semin. Thromb. He- 42

~~

mosthsis 5,'254-273

(1980) Nature 286, 456-460

(1982) J. Bwchem. (Tokyo) 92,689-698

3. Heimark, R. L., Kurachi, K., Fujikawa, K., and Davie, E. W.

4. Sugo, T., Hamaguchi, A., Shimada, T., Kato, H., and Iwanaga, S.

5. Griffin, J. H., and Cochrane, C. G. (1976) Proc. Natl. Acad. Sci. U. S. A. 73,2554-2558

6. Meier, H. L., Pierce, J. V., Colman, R. W., and Kaplan, A. P.

7. Ichinose, A., Fujikawa, K., and Suyama, T. (1986) J. Bwl. Chem.

8. Zimmerman, T. S., Fierer, J., and Rothberger, H. (1977) Semin.

(1977) J. Clin. Inuest. 60, 18-31

261,3486-3489

Hemafol. 14, 391-408

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31. 32.

33.

34.

35.

Mandle, R. J., Colman, R. W., and Kaplan, A. (1976) Proc. Natl.

Thompson, R. E., Mandle, R., and Kaplan, A. P. (1977) J. Clin.

Wiggins, R. C., Bouma, B. N., Cochrane, C. G., and Griffin, J. H.

Thompson, R. E., Mandle, R., and Kaplan, A. P. (1979) Proc.

Kerbiriou, D. M., Bouma, B. N., and Griffin, J. H. (1980) J. Bid.

Bock, P. E., Shore, J. D., Tans, G., and Griffin, J. H. (1985) J.

Acad. Sci. U. S. A. 73,4179-4183

Znuest. 60,1376-1380

(1977) Proc. Natl. Acad. Sci. U. S. A. 74, 4636-4640

Natl. Acad. Sci. U. S. A. 76, 4862-4866

Chem. 235,3952-3958

Biol. Chem. 260. 12434-12443 Bouma, B.~N., Vlodswijk,R. A. A., and Griffin, J. H. (1983) Blood

Warn-Cramer, B. J., and Bajaj, S. P. (1985) Biochem. Bwphys. 62, 1123-1131

Res. Commun. 133,417-422 Bock, P. E., and Shore, J. D. (1983) J. Biol. Chem. 258, 15079-

Shimada, T., Kato, H., Maeda, H., and Iwanaga, S. (1985) J.

Ikari, N., Sugo, T., Fujii, S., Kato, H., and Iwanaga, S. (1981) J.

Fujikawa, K., and McMullen, B. A. (1983) J. Biol. Chem. 258,

McMullen, B. A,, and Fujikawa, K. (1985) J. Bwl. Chem. 260,

Chung, D. W., Fujikawa, K., McMullen, B. A., and Davie, E. W.

Fujikawa, K., Chung, D. W., Hendrickson, L. E., and Davie, E.

Lottspeich, F., Kellerman, J., Henschen, A., Foertsch, B., and

Takagaki, Y., Kitamura, N., and Nakanishi, S. (1985) J. Biol.

Kellerman, J., Lottspeich, F., Henschen, A., and Muller-Esterl,

Tait, J. F., and Fujikawa, K. (1986) J. Bid. Chem. 261, 15396-

Heimark, R. L., and Davie, E. W. (1981) Methods Enzymol. 80,

Kurachi, K., and Davie, E. W. (1977) Biochemistry 16, 5831-

15086

Biochem. (Tokyo) 97, 1637-1644

Biochem. (Tokyo) 89,1699-1709

10924-10933

5328-5341

(1986) Biochemistry 25,2410-2417

W. (1986) Biochemistry 25, 2417-2424

Muller-Esterl, W. (1985) Eur. J . Biochem. 152, 307-314

Chem. 260,8601-8609

W. (1986) Eur. J. Biochem. 154, 471-478

15401

157-172

5829 Kato, H., Nagasawa, S., and Iwanaga, S. (1981) Methods Enzymol. ""

80, 172-198 Steinbach, G. (1974) Acta Histochem. 50, 19-34 Fujikawa, K., Thompson, A. R., Legaz, M. E., Meyer, R. G., and

Lakowicz, J. R. (1983) Principles of Fluorescence Spectroscopy,

Dandliker, W. B., Kelly, R. J., Dandliker, J., Farquhar, J., and

van der Graaf, F., Tans, G., Bouma, B. N., and Griffin, J. H.

Davie, E. W. (1973) Biochemistry 12,4938-4945

Plenum Publishing Corp., New York

Levin, J. (1973) Immunochemistry 10,219-227

(1982) J. Bwl. Chem. 267. 14300-14305 36. Scott, C. F., Schapira, M.,- James, H. L., Cohen, A. B., and

Colman, R. W. (1982) J. Clin. Znuest. 69, 844-852 37. Colman, R. W. (1984) J. Clin. Inuest. 73, 1249-1253 38. Lewis, G. P. (1970) in Handbook of Experimental Pharmacology

(Erdos, E. G., ed) Vol. 25, pp. 516-530, Springer-Verlag New York Inc., New York

39. Schapira, M., Silver, L. D., Scott, C. F., Schmaier, A. H., Prograis, L. J.. Curd. J. G.. and Colman, R. W. (1983) N . E n d J. M d

\"", ~ ~ ~- ~ ~ ~

308; 105011053 ' 40. Baumgarten, C. R., Togias, A. G., Naclerio, R. M., Lichtenstein,

L. M., Norman, P. S., and Proud, D. (1985) J. Clin. Znuest. 76. 191-197

41. Schapira, M., Despland, E., Scott, C. F., Boxer, L. A., and Colman, R. W. (1982) J. Clin. Znuest. 69, 1199-1202

42. Wachtfogel, Y. T., Kucich, U., James, H. L., Scott, C. F., Schap- ira, M., Zimmerman, M., Cohen, A. B., and Colman, R. W. (1983) J. Clin. Znuest. 72, 1672-1677

43. Iwanaga, S., Kato, H., Sugo, T., Ikari, N., Hashimoto, N., and Fujii, S. (1979) in Biological Functions of Proteinases (Holzer, H., and Tschesche, H., eds) pp. 243-259, Springer-Verlag, Hei- delberg

44. Kitamura, N., Takagaki, Y., Furuto, S., Tanaka, T., Nawa, H., and Nakanishi, S. (1983) Nature 305,545-549

45. Weber, K., and Osborn, M. (1969) J. Biol. Chem. 244,4406-4412