THE JOURNAL OF BIOLOGICAL CHEMISTRY 8 1985 by The American Society of Bioiogical Chemists, Inc.

Val. 260, No. 6, Issue of March 25, pp. 3501-3505, 1985 Printed in U.S.A.

The Protease Specificity of Heparin Cofactor 11 INHIBITION OF THROMBIN GENERATED DURING COAGULATION*

(Received for publication, September 24,1984)

Katherine A. Parkerg and Douglas M. Tollefsen$? From the Division of Hematology-Oncology, Departments of internal Medicine and Biological Chemistry, Washington University School of Medicine, St. Louis, Missouri 631 10

'281-labeled heparin cofactor I1 (HCII) was mixed with plasma and coagulation was initiated by addition of CaClz, phospholipids, and kaolin or tissue factor. In the presence of 67 ggfml of dermatan sulfate, radioac- tivity was detected in a band which corresponded to the thrombin-HCII complex (M. = 96,000) upon sodium dodecyl sulfate-polyacrylamide gel electrophoresis. No other complexes were observed. The thrombin-HCII complex was undetectable when 5 units/ml of heparin was present or when prothrombin-deficient plasma was used. In experiments with purified proteases, HCII did not significantly inhibit coagulation factors VIIa, IXa, Xa, XIa, XIIa, kallikrein, activated protein C, plasmin, urokinase, tissue plasminogen activator, leu- kocyte elastase, the r-subunit of nerve growth factor, and the epidermal growth factor-binding protein. HCII inhibited leukocyte cathepsin G slowly, with a rate constant of 8 X lo4 I"1 min" in the presence of der- matan sulfate. These results indicate that the protease specificity of HCII is more restricted than that of other plasma protease inhibitors and suggest that the anti- coagulant effect of dermatan sulfate is due solely to inhibition of thrombin by HCII.

"

Heparin cofactor I1 (HCII') is a 65,600-dalton glycoprotein in human plasma which inhibits thrombin by forming a stable, equimolar complex with the protease (1-7). Heparin and dermatan sulfate bind to HCII and thereby increase the rate of inhibition of thrombin -1000-fold (4, 5, 8). Heparin also catalyzes the inhibition of thrombin and other proteases by antithrombin I11 (ATIII) (9). In contrast, dermatan sulfate specifically catalyzes the thrombin-HCII reaction but has no appreciable effect on the activity of ATIII (5, 10). HCII is present in plasma at a concentration of -1.2 P M . ~ At this concentration, thrombin could theoretically be inhibited with a t I j 2 approaching 50-100 ms in the presence of an optimal

* This research was supported by grants from the National Insti- tutes of Health (HL-27589) and the Monsanto Co. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "adver- tisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

4 Present address: Department of Internal Medicine, University of Texas Southwestern Medical School, Dallas, TX.

Award HL-01079. 8 Recipient of National Institutes of Health Career Development

'The abbreviations used are: HCII, heparin cofactor 11; ATIII, antithrombin 111; TPA, tissue plasminogen activator; yNGF, the pept,idase subunit of nerve growth factor; EGF-BP, epidermal growth factor-binding protein, the peptidase subunit of epidermal growth factor; and SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

* D. M. Tollefsen and C. A. Pestka, manuscript submitted.

amount of heparin or dermatan sulfate. Little information is available about the protease specificity

of HCII. In previous studies, HCII did not inhibit coagulation factor Xa, plasmin, or trypsin (2, 4, 6, 7). In contrast, ATIII inhibits a broad range of proteolytic enzymes, including the coagulation factors thrombin, Xa, IXa, XIa, XIIa, and kalli- krein, and the fibrinolytic enzyme plasmin (9, 11). We have now examined all of the proteases known to be involved in coagulation and fibrinolysis, as well as several other extracel- lular proteases, and have discovered that HCII is a relatively specific inhibitor of thrombin. Previously we reported that ' " ~ - t ~ r o ~ b i n added to plasma containing dermatan sulfate becomes bound to HCII (5). We have now shown that throm- bin generated in plasma during coagulation is inhibited by HCII when dennatan sulfate is present, thus explaining the anticoagu~ant effect of dermatan sulfate that has been ob- served in vitro (10, 12).

EXPERIMENTAL PROCEDURES

Materiok-Benzoyl-11-Glu-Gly-Arg-pnitroanilide (S-22221,pyro- Glu-Gly-Arg-p-nitroanilide (5-24441, Val-Leu-Lys-p-nitroanilide (S- 2251), Pro-Phe-Arg-p-nitroanilide (8-23023, and Phe-Pip-Arg-p-ni- troanilide (5-2238) were purchased from Helena Laboratories; succi- nyl-Ala-Ala-Pro-Phe-p-nitroanilide from Vega Biochemicals; succi- nyl-A~a-Ala-Ala-p-nitroanilide from Sigma; and tosyl-Gly-Pro-Arg- p-nitroanilide (Chromozyme TH) from Boehringer Mannheim. Hep- arin from porcine intestinal mucosa was obtained from Abbott Lab- oratories. Porcine skin dermatan sulfate was obtained from Sigma and was treated with nitrous acid prior to use to remove contaminat- ing heparin (5, 10). S ~ i ~ m [ ~ * ~ I ~ i o d i d e (16.8 Ci/mg) was purchased from Amersham. Iodogen was purchased from Pierce. Prothrombin- deficient plasma containing -2% of the normal concentration of prothrombin was purchased from George King Biologicals. Normal plasma was obtained from blood (4.5 ml) drawn into evacuated tubes containing 0.5 mi of 0.129 M buffered sodium citrate (Vacutainer #6418, Bectin-Dickinson). Activated partial thromboplastin reagent was obtained from Hyland Laboratories. Rabbit brain throm~plast in was obtained from Ortho Diagnostics and was reconstituted with water according to the manufacturer. Human brain thromboplastin (13) was obtained from Dr. George Broze, Washington University. Rabbit brain cephalin was purchased from Sigma. Polybrene (1,5- dimethyl-l,5-~azaundecamethyIene polymethobromidef was ob- tained from Aldrich.

Proteins-Human HCII and thrombin were purified as previously described (4). Human factor XIIa (14) was prepared by Dr. Allen Kaplan, State University of New York, Stony Brook. Human factor XIa (15) was obtained from Dr. Paul Bajaj, University of California, San Diego. Human coagulation factors VIIa (13), IX, X, Xa (16), and activated protein C (17) were obtained from Drs. Hatem Salem, George Broze, and Joseph Miletich, ~ a s h i n ~ o n University. Factor IX (69 FM) was converted to IXa by incubation with 40 nM factor XIa for 2 h at 37 "c in buffer containing 5 mM CaC12, 0.05 M NaCl, 0.02 M Tris-HC1, pH 7. Tissue plasminogen activator derived from cultured human melanoma cells (18) was obtained from Dr. DesirB Collen, University of Leuven. Human urokinase was the product of Winthrop Laboratories. Glu-plasminogen I1 was purified from human plasma by the method of Deutsch and Mertz (19) and the zymogen

3501

3502 Protease Specificity of HCII (1 p ~ ) was converted to plasmin by incubation with 50 nM urokinase for 30 min at 37 "c in buffer containing 8 mM lysine, 20% glycerol, 0.8 M Tris-HC1, pH 9.6. Human leukocyte elastase (20) and cathepsin G (21, 22) were prepared by Dr. Robert Senior, Washington Univer- sity. The proteolytic subunits of murine nerve growth factor (-y-NGF) (23) and epidermal growth factor (EGF-BP) (24) were provided by Dr. Thomas Maciag, Meloy Laboratories. Human plasma kallikrein, human fibrinogen, Bothrops atrox venom, and bovine serum albumin were purchased from Sigma.

Iodination of HCII-200 pl of HCII (-8 p~ in 0.15 M NaC1, 0.02 M Tris-HC1, pH 7.4) were incubated with 2 mCi of carrier-free NaIZ5I for 10 min at 4 "C in a polypropylene tube coated with 100 pg of Iodogen according to the method of Fraker and Speck (25). Unbound Iz5I was removed by gel filtration on a 1 X 10-cm column of Sephadex (2-25. The final specific radioactivity was 1.2 X lo9 cpm/nmol of HCII.

Electrophoresis-SDS-PAGE was performed with 7.5% gels and the Laemmli buffer system (26) under nonreducing conditions. Au- toradiography was performed as described previously (3). Molecular weight standards obtained from BioRad included ovalbumin (M, = 45,000), bovine serum albumin (M, = 66,000), phosphorylase b (M, = 92,000), @-galactosidase (M, = 116,000), and myosin (M, = 200,000).

Assays for Thrombin, Factors Xa, XIa, Kallikrein, Actiuated Protein C, Plasmin, Urokinase, Cathepsin G, Elastase, y-NGF, and EGF-BP- Reactions were carried out in 1.5-ml polypropylene microcentrifuge tubes at 37 "C. The enzyme, HCII, heparin, and dermatan sulfate were mixed at the final concentrations specified in Table I in buffer containing0.15 M NaCl, 0.02 M Tris-HC1, pH 7.4, and 1 mg/ml bovine serum albumin. Reactions were initiated by addition of the protease. At specified times, protease activity was determined by addition of an equal volume of the appropriate chromogenic substrate in water. The rate of hydrolysis of the substrate was determined from a continuous recording of the absorbance at 405 nm. Alternatively, hydrolysis of the substrate was terminated after 3 min by addition of 1/10 volume of 50% (v/v) acetic acid, the precipitated glycosamino- glycan was removed by centrifugation if necessary, and the absor- bance of the supernatant solution was determined at 405 nm. In the absence of HCII, a standard curve of AA405 versus protease concen- tration was linear for each protease in the concentration range employed in the experiment. The following substrates were used 0.1 mM Chromozym T H for thrombin, r-NGF, and EGF-BP 0.3 mM S- 2222 for factor Xa; 3 mM S-2222 for factor XIa; 0.4 mM S-2302 for kallikrein; 0.2 mM S-2238 for activated protein C; 0.5 m M S-2251 for plasmin; 0.6 mM $2444 for urokinase; 0.2 mM succinyl-Ala-Ala-Pro- Phe-p-nitroanilide for cathepsin G; and 0.2 mM succinyl-Ala-Ala- Ala-p-nitroanilide for elastase.

Assay for factor VIIa-Factor VIIa was incubated with HCII and heparin or dermatan sulfate in the presence of 5 mM CaClz and 1 mg/ ml human brain thromboplastin in 0.15 M NaC1, 0.02 M Tris-HC1, pH 7.4, for 15 min at 37 "C. Tritiated human factor X (4.8 X lo6 cpm/nmol) prepared by the method of Silverberg et al. (27) was then added at a final concentration of 0.4 p ~ . After 1 min, proteolysis of 3H-factor X was terminated by addition of 50 mM EDTA, and the protein was precipitated with 5% trichloroacetic acid at 4 "C. The radioactivity of the supernatant solution containing the labeled acti- vation peptide of factor Xa was then determined. In control incuba- tions without HCII, the amount of 3H-peptide released was propor- tional to the concentration of factor VII..

Assay for Factor IXa-Factor IXa, HCII, and heparin or dermatan sulfate were incubated for 15 min at 37 "C in 55 pl of buffer containing 0.15 M NaCI, 0.02 M Tris-HC1, pH 7.4, and 1 mg/ml bovine serum albumin. At the end of the incubation period, 45 pl of 0.1 mg/ml Polybrene in the above buffer, 100 pl of citrate-anticoagulated normal plasma, 50 pl of rabbit brain cephalin, and 100 pl of 25 mM CaClZ were added sequentially a t 15-5 intervals. The clotting time following the addition of CaC& was determined with a Fibrometer (Bectin- Dickinson). Samples (55 pl) of factor IXa at concentrations of 15- 245 nM yielded clotting times which decreased from 72 to 36 s. A plot of log clotting time versus log [factor IXa] was linear in this range.

Assay for Factor XIIa-Factor XIIa was assayed in the laboratory of Dr. Allen Kaplan as previously described (14).

Assay for Tissue Plasminogen Actiuator-TPA, HCII, and heparin or dermatan sulfate were mixed with 0.9 mg/ml fibrinogen in 0.15 M NaC1, 0.02 M Tris-HC1, pH 7.4. B. atrox venom (1 unit/ml final concentration) was added immediately to clot the fibrinogen. After a 60-min incubation at 37 "c, an equal volume of 0.6 mM $3-2444 was added. Hydrolysis of the substrate was terminated after 3 min by addition of acetic acid (5% final concentration), the mixture was centrifuged for 10 min in an Eppendorf microcentrifuge, and the absorbance at 405 nm of the supernatant solution was determined. Control experiments demonstrated (a) that AM5 was proportional to the amount of TPA present in the absence of HCII; (b) that the venom did not hydrolyze the substrate S-2444; and (c) that HCII did not inhibit nor was it degraded by the venom protease. In preliminary experiments, fibrinogen and B. atrox venom were omitted from con- trol incubations without HCII; as previously reported (18), the rate of substrate hydrolysis by TPA in the absence of fibrin was -30% of the rate obtained in the presence of fibrin.

Kinetic Analysis-The second-order rate constant for inhibition of a protease by HCII was estimated according to the equation: k = ln([P]o/[P],)/t/[[HCII], in which [PIo = initial protease activity, [PIl

TABLE I Inhibition of purified proteases by HCII

Reactants were incubated at 37 "C at the final concentrations indicated in 0.15 M NaCl, 0.02 M Tris-HC1, and 1 mg/ml bovine serum albumin, pH 7.4. At the end of the incubation period, the remaining protease activity was determined as described under "Experimental Procedures."

Activity remaining after incubation with

Protease [Protease] [HCII] Incubation [Heparin] 'Dermatan sulfate] HCII Heparin HCII + Dermatan HC1l +

alone sulfate time

alone alone heparin

nM min unitlrnl ccglml % of control' Thrombin 2.2 150 20 0.5 100 22 101 <2 100 <2 Factor Xa 11.6 200 30 1 100 109 NDb 94 ND 108 Factor VIIa 0.83 760 15 1 80 108 ND 110 ND 115 Factor IXa 69 1220 20 1 100 95 ND 106 ND 116 Factor XIa 2.8 100 15 1 100 105 103 106 79 83 Factor XIIa 4.3 50 15 1 100 100 101 92 98 95 Kallikrein 2.9 100 10 1 100 102 98 99 100 101 Protein Ca 83 760 10 1 100 98 98 99 47 50 Plasmin 290 750 30 1 100 100 101 84 87 98 Urokinase 540 830 30 1 100 107 108 104 109 106 TPA 55 760 60 0.5 400 102 51 56 95 95 Cathepsin G 730 1250 40 0.5 400 30 43 36 76 <2 Elastase 670 1780 60 0.5 400 56 45 50 95 57 y-NGF 100 760 10 1 100 106 105 103 91 91 EGF-BP 100 760 10 1 100 95 95 96 76 69

a Averages of duplicate determinations. ND, not determined.

Protease Specificity of HCII 3503

= protease activity at time = t, and [HCII] = initial HCII concentra- tion, assuming pseudo-first order conditions.

RESULTS

Inhibition of Thrombin by HCII in Recalcified Plasma-We have previously shown that thrombin combines with HCII to form a 96,000-dalton complex that is stable during SDS- PAGE (3, 4). To determine whether HCII inhibits thrombin or other proteases as they are generated during coagulation, tracer '251-HCII was added to citrate-anticoagulated plasma, the mixture was warmed to 37 "C in a glass tube, and coagu- lation was initiated by addition of CaC12. After 1 h, the plasma was analyzed by SDS-PAGE and autoradiography (Fig. 1, lanes E-G). Coagulation occurred in the absence of heparin or dermatan sulfate, and none of the lZ5I-HCII was detected in higher molecular weight complexes (lune E). However, coagulation did not occur in the presence of 67 pg/ml of dermatan sulfate, and densitometry revealed that -13% of the label was present in a 96,000-dalton band (lane G) which co-migrated with the complex formed by incubation of an excess of purified thrombin with '251-HCII (lune A ) . Less than 1% of the label was present in the complex when prothrombin- deficient plasma was substituted for normal plasma in the incubation (not shown). Coagulation also did not occur in the presence of 5 units/ml of heparin. In this case, none of the '251-HCII was detected in the 96,000-dalton complex (lane F). This result is consistent with previous experiments which demonstrated that '251-thrombin is preferentially inhibited by ATIII in undiluted plasma at similar concentrations of hep- arin (3). Control incubations from which the CaC12 was omit-

200 -

116 - 921

66 -

45 - MW Stds.

FIG.

A B C D E F G H I J K L M

1. IncorDoration of 12'I-HCII into comdexes in Dlasma during coagulation. Reagents were brought t o a final voiume of 150 pl with 0.15 M NaC1, 0.02 M Tris-HC1, 1 mg/ml bovine serum albumin, pH 7.4, and incubated for 1 h at 37 "C in glass (lunes B-G) or polypropylene (lunes A and H-M) tubes. Each incubation con- tained 7.5 nM lZ5I-HCII along with the following additional reagents: lune A , 13 p~ thrombin; lane B, 25 p1 of normal plasma; lune C, same as I3 plus 5 units/ml heparin; lune D, same as B plus 67 pg/ml dermatan sulfate; lune E, 25 pl of normal plasma and 6 mM CaCl2; lune F, same as E plus 5 units/ml heparin; lune G, same as E plus 67 pg/ml dermatan sulfate; lune H , 25 pl of prothrombin-deficient plasma, 13 mM CaClZ, and 100 pl of human brain thromboplastin; lune I, same as H plus 5 units/ml heparin; lane J , same as H plus 67 pg/ml dermatan sulfate; lune K, 25 p1 prothrombin-deficient plasma, 8 mM CaClZ, and 50 p1 of activated partial thromboplastin reagent; lune L, same as K plus 5 units/ml heparin; lune M , same as K plus 67 pg/ml dermatan sulfate. At the end of the incubation period, 5-10 p1 of each reaction mixture were subjected to SDS-PAGE. An autora- diograph of the gel exposed for 1 h at -70 "C is shown. The positions of molecular weight standards and of the 96,000-dalton thrombin- HCII complex (+) are indicated.

ted did not contain labeled complexes regardless of whether heparin or dermatan sulfate was present (lanes B-D). These experiments indicate that thrombin generated in plasma by activation of the intrinsic coagulation pathway is inhibited by HCII in the presence of dermatan sulfate.

Inhibition of Thrombin by HCII in Plasma Activated by Tissue Factor or Kaolin-In an attempt to detect complexes of HCII with proteases other than thrombin, prothrombin- deficient plasma containing '251-HCII was incubated for 1 h at 37 "C with CaC12, phospholipids, and a source of tissue factor (human brain thromboplastin) to activate factor VI1 or kaolin (activated partial thromboplastin reagent) to activate factor XI1 and kallikrein. In neither case was any of the lZ5I- HCII detected in complexes (Fig. 1, lanes Hand K). Identical results were obtained in the presence of 5 units/ml of heparin (lanes I and L). When incubations were repeated in the presence of 67 pg/ml of dermatan sulfate, a trace amount of the label was present in a band corresponding to the throm- bin-HCII complex (lanes J and M). No other complexes were observed. When the exposure times of the autoradiographs in Fig. 1 were extended from 1 to 20 h to increase the sensitivity of the experiments, we observed numerous additional bands representing as a whole 0.1% of the total radioactivity present. Because there were no significant differences between the additional bands and the pattern of a gel containing '251-HCII alone (not shown), the bands were considered to represent trace contaminants in the HCII preparation.

Inhibition of Purified Proteases by HCII-We assayed four- teen purified proteases for activity after incubation with a molar excess of HCII (Table I). The concentrations of pro- tease and HCII were determined primarily according to the sensitivity of the assay for the protease. Incubation times were long in comparison to the t112 for inhibition of thrombin by HCII (e.g. tl,2 = 8 s in the presence of 50 nM HCII and 0.5 unit/ml of heparin; Ref. 4). Under the conditions of each experiment, 20% inhibition of the protease would indicate a second-order rate constant 1 3 X lo5 M" min" (ie. 2000 times less than the rate constant for inhibition of thrombin by HCII in the presence of dermatan sulfate; see "Discussion"). Hep- arin or dermatan sulfate were present at concentrations pre- viously determined to accelerate the inhibition of thrombin by HCII (4 ,5) . In addition, controls were performed to deter- mine the effects of heparin and dermatan sulfate alone on protease activity. As shown in Table I, 22% of thrombin activity remained after a 20-min incubation with 150 mM HCII, while <2% activity remained in incubations that also included heparin or dermatan sulfate. The second-order rate constant calculated for inhibition of thrombin by HCII alone in this experiment was 5 X lo5 M" min", as previously reported (4). Rate constants for inhibition of thrombin by HCII in the presence of heparin or dermatan sulfate could not be determined accurately from the data in Table I, but in both cases were >1.3 x lo6 M" min" (see "Discussion"). In contrast, HCII did not inhibit significantly coagulation factors VIIa, IXa, Xa, XIa, XIIa, kallikrein, activated protein C, plasmin, urokinase, TPA? or 7-NGF. Activated protein C and TPA were moderately inhibited by dermatan sulfate and heparin, respectively, but there was no further inhibition in either case when HCII was also present. Leukocyte elastase was partially inhibited during a 60-min incubation with 1.78 p~ HCII alone or with dermatan sulfate; in both cases the rate constants were 5 x lo3 M" min". Although elastase was moderately inhibited by heparin alone, heparin appeared to

HCII was incubated with TPA alone (data not shown) or in the presence of fibrin as described under "Experimental Procedures" (data in Table I).

3504 Protease ~ ~ e ~ ~ f i c i t ~ of He11

protect the protease from further inhibition by HCII. EGF- BP was inhibited partially by dermatan sulfate and to a slightly greater extent when both dermatan sulfate and HCII were present; in the latter case the calculated rate constant was 1 X io4 M" min-'.

Inhibition of Cathepsin G by HCZZ-Because HCII appeared to inhibit cathepsin G at a significant rate in the presence of dermatan sulfate (Table I), the time course of inhibition was studied in more detail (Fig. 2). In the absence of heparin or dermatan sulfate, inhibition by 1.25 PM HCII occurred with a tllz = 24 min ( k = 1.4 X lo4 M-' min-I). Although dermatan sulfate alone decreased cathepsin G activity 20-30% (Table I), it increased the rate of inhibition by HCII &fold (tIl2 = 4 min; k = 8.4 x IO5 M" min"). In contrast, heparin decreased the activity of cathepsin G 50-60%, and it appeared to prevent further inhibition of the protease by HCII.

DISCUSSION

The purpose of this investigation was to identify enzymes that HCII can inhibit, and thereby to arrive at a hypothesis concerning the physiological function of HCII. We have tested various serine proteases, including all of those currently known to be involved in coagulation and fibrinolysis (28), leukocyte cathepsin G (21) and elastase (ZO), and the pepti- dase subunits of nerve growth factor (23) (y-NGF) and epi- dermal growth factor (24) (EGF-BP). Of the enzymes tested

0.50 r 0,40\ 0.3oJi \

0

m 0 w z

z 0.10 -

2 0.08 .. -

U - 0.06 -

r 10 20 30 4 0

FIG. 2. Inhibition of leukocyte cathepsin G by HCII. Cathep- sin G (0.73 p ~ ) and HCII (1.25 p ~ ) were incubated for 1-40 min at 37 "C with 0.5 unit/ml heparin (O), 100 pg/ml dermatan sulfate (A), or no glycosaminoglycan (0) in 33 pl of 0.15 M NaCl, 0.02 M Tris- HC1, 1 mg/ml bovine serum albumin, pH 7.4. Then 167 pl of 0.35 mg/ml succinyl-Ala-Ala-Pro-Phe-p-nitroanilide in water was added. Hydrolysis of the substrate was stopped after 3 min by addition of 20 pl of 50% (v/v) acetic acid, the mixture was centrifuged for 10 min in an Eppendorf microcentrifuge, and the absorbance of the supernatant solution was determined at 405 nm. In control experiments the absorbance was directly p r o ~ ~ i o n a l to the concentration of enzyme added in the absence of HCII and glycosaminoglycan.

TIME (Min)

other than thrombin, only leukocyte cathepsin G was inhib- ited at a significant rate. However, the calculated second- order rate constant for inhibition of cathepsin G by HCII in the presence of dermatan sulfate was -40,000-fold less than the rate constant reported for inhibition of cathepsin G by a1-antichymotrypsin (29). Therfore, inhibition of cathepsin G by HCII is unlikely to occur i n uiuo. In contrast, HCII inhibits thrombin with rate constants of 6.4 X lo8 M" min" in the presence of 250 pg/ml of dermatan sulfate and 4.0 X 10' M" min" in the presence of 10 units/ml of heparin (5). Rate constants of this magnitude are characteristic of inhi- bition reactions that are likely to be "physiological" (30).

Our data indicate that the protease specificity of HCII is more restricted than that of other plasma protease inhibitors, including ATIII (9), al-proteinase inhibitor (30), a2-antiplas- min (31), and a2-macroglobulin (32), each of which can inhibit several of the proteases that we have tested. In addition, the lack of inhibition of y-NGF, EGF-BP, plasmin, and urokinase distinguishes HCII from the cellular protease inhibitors termed "protease nexins" (33). We have also found that, in the presence of dermatan sulfate, HCII binds thrombin as it is being generated in plasma during coagulation. Thus, inhi- bition of thrombin by HCII appears to explain the anticoag- ulant activity of dermatan sulfate that has been observed in vitro (lo, 12) and may also explain the antithrombotic effect observed in uiuo after the administration of exogenous der- matan sulfate (34). In addition, HCII may inhibit other effects of thrombin, including platelet aggregation and secretion (35), chemotaxis (36), and mitogenesis (37), under appropriate circumstances.

Rapid inhibition of thrombin by HCII i n vivo probably occurs only in the immediate vicinity of proteoglycans which contain oligosaccharide sequences that bind HCII (4,8). Sim- ilarly, ATIII requires specific oligosaccharide sequences for maximum activity (9). HCII and ATIII may become activated in different environments, since there is evidence that differ- ent heparin molecules activate HCII and ATIII (38) and that dermatan sulfate only activates HCII (5). Dermatan sulfate comprises 60-70% of the glycosaminoglycans in the intima and media of large arteries (39), in addition to being present in skin, heart valves, and tendons (40). A small amount of dermatan sulfate is also synthesized by cultured endothelial cells (41). Whether the glycosaminoglycans present in these locations contain the proper sequences to active HCII remains to be determined.

Acknowledgments-We thank Dr. Allen Kaplan for performing the factor XIIa assays and Drs. Paul Bajaj, George Broze, DBsirB Collen, Thomas Maciag, Joseph Miletich, Hatem Salem, and Robert Senior for providing proteases and other reagents.

REFERENCES 1. Briginshaw, G. F., and Shanberge, J. N. (1974) Arch. Biochem.

2. Briginshaw, G. F., and Shanberge, J. X. (1974) Thromb. Res. 4,

3. Tollefsen, D. M., and Blank, M. K. (1981) J. Clin. Invest. 68,

4. Tollefsen, D. M., Majerus, D. W., and Blank, M. K. (1982) J.

5. Tollefsen, D. M., Pestka, C. A., and Monafo, W. J. (1983) J. Biol.

6. Griffith. M. J.. Carrawav, T., White, G. C., and Dombrose, F. A.

Biophys. 161,683-690

463-477

589-596

Rid. Chem. 257,2162-2169

Chem. 258,6713-6716

(1983) mood 61,111-118 7. Wundenvald. P.. Schrenk. W. J.. and Port, H. (1982) Thromb.

Res. 25, 177-191

Commun. 112,663-670 8. Griffith, M. J., and Marbet, G. A. (1983) Biochem. Biophys. Res.

9. Rosenberg, R. D. (1977) Semin. Hematol. 14,427-440 10. Teien, A. X., Abildgaard, U., and Hook, M. (1976) Thromb. Res.

Protease Specificity of HCII 3505

8,859-867 11. Highsmith, R. F., and Rosenberg, R. D. (1974) J. Bioi Chem.

12. Long, W. F., Williamson, F. B., Kindness, G., and Edward, M.

13. Broze, G. J., and Majerus, P. W. (1980) J. Bid. Chem. 255,

14. Silverberg, M., Dunn, J. T., Garen, L., and Kaplan, A. P. (1980)

15. Bajaj, S. P. (1982) J. Biol. Chem. 257,4127-4132 16. Miletich, J. P., Broze, G. J., and Majerus, P. W. (1980) Anal.

17. Salem, H. H., Broze, G. J., Miletich, J. P., and Majerus, P. W.

18. Rijken, D. C . , and Collen, D. (1981) J. Biol. Chem. 256, 7035-

19, Deutsch, D. G., and Mertz, E. T. (1970) Science 170,1095-1096 20. Baugh, R, J., and Travis, J. (1976) biochemist^ 15,836-841 21. Barrett, A. J. (1981) ~ e t ~ o ~ ~ n z y m o ~ . 80,561-565 22. Senior, R. M., and Campbell, E. J. (1984) J. Immunol. 132,

23. Greene, L. A,, Shooter, E. M., and Varon, S. (1969) B i o c ~ m ~ ~ ~

24. Taylor, J. M., Mitchell, W. M., and Cohen, S. (1974) J. Biol.

25. Fraker, P. J., and Speck, J. C., Jr. (1978) Biochem. Biophys. Res.

26. Laemmli, U. K. (1970) Nature (Land.) 227, 680-685

249,4335-4338

(1980) Thromb. Res. 18, 493-503

1242-1247

J. Biol. Chem. 255, 7281-7286

Biochem. 105, 304-310

(1983) Proc. Natl. Acad. Sci. U. S. A. 80, 1584-1588

7041

2547-2551

8,3735-3741

Chem. 249,2188-2194

Commun. 80, 849-857

27. Silverberg, S. A., Nemerson, Y., and Zur, M. (1977) J. Biol. Chem.

28. Jackson, C. M., and Nemerson, Y. (1980) Annu. Reu. Biochem.

29. Beatty, K., Bieth, J., and Travis, J. (1980) J. Biol. Chem. 255,

30. Travis, J., and Salvesen, G. S. (1983) Annu. Reu. Biochem. 52,

31. Saito, H., Goldsmith, G. H., Moroi, M., and Aoki, N. (1979) Proc.

32. Roberts, R. C., and Hall, P. K. (1983) Ann. N . Y. Acad. Sci. 421,

33. Knauer, D. J., Scaparro, K. M., and Cunningham, D. D. (1982)

34. Buchanan, M. R., Boneu, B., Ofosu, F., and Hirsh, J. (1985) Blood

35. Sie, P., Fernandez, F., Caranobe, C., Gabaig, A. M., and Boneu,

36. Bar-Shavit, R., Kahn, A., Mudd, M. S., Wilner, G. D., Mann, K.

252,8481-8488

49,765-811

3931-3934

655-709

Natl. Acad. Sci. U. S. A. 76,2013-2017

6 1-68

J. Biol. Chem. 257, 15098-15104

65,198-201

B. (1984) Thromb. Res. 35, 231-236

G., and Fenton, J. W., I1 (1984) B i o c h e m ~ t ~ 23,397-400 37. Chen, L. B., and Buchanan, J. M. (1975) Pro;. Natl. Acad. Sci.

U. S. A. 72, 131-135 38. Hurst, R. E., Poon, M.-C., and Griffith, M. J. (1983) J. Clin.

39. Wight, T. N., and Ross, R. (1975) J. Cell Bioi 67,675-686 40. Kumar, V., Berenson, G. S., Ruiz, H., Dalferes, E. R., and Strong,

41. Oohira, A., Wight, T. N., and Bornstein, P. (1983) J. Biol. Chem.

Invest. 72,1042-1045

J. P. (1967) J. Atheroscler. Res. 7,583-590

258,2014-2021