THEJOURNAL OF BIOLOGICAL CHEMISTRY Vol. 267, No. 13, Issue of May 5, pp. 8789-8794, 1992 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S. A.

Heparin Binding toProtein C Inhibitor*

(Received for publication, December 11, 1991)

Charlotte W. PrattS and FrankC. Church From the DeDartment of Pathology and The Center for Thrombosis and Hemostasis, The University of North Carolina School of Medicine, C&el Hill, North Ca&na 27599

Protein C inhibitor isa plasma protein whose ability regulated by a plasma glycoprotein named protein C inhibitor, to inhibit activated protein C, thrombin, and other also known as plasminogen activator inhibitor-3 (4). Three enzymes is stimulated by heparin. These studies were other major plasma proteins, az-macroglobulin, az-antiplas- undertaken to further understand how heparin binds min, and a’-proteinase inhibitor inhibit activated protein C to protein C inhibitor andhow it accelerates proteinase (5) but might be effective only by virtue of their relatively inhibition. The region of protein C inhibitor from res- high concentration in plasma. Protein C inhibitor reacts with idues 264-283 was identified as the heparin-binding the active site of activated protein C to form an essentially site. This differs from the putative heparin-binding irreversible complex (6). Interestingly, protein C inhibitor site in the related proteins antithrombin and heparin also inhibits thrombin, the final proteinase of the coagulation cofactor. The glycosaminoglycan specificity of protein pathway, as well as other procoagulant enzymes. This broad C inhibitor was relatively broad,including heparin and heparan sulfate, but not dermatan sulfate. Non-sul- target proteinase specificity of protein C inhibitor presents a fated and non-carboxylated polyanions also enhanced problem in understanding the physiological importance of proteinase inhibition by protein C inhibitor. Heparin protein C inhibitor as a regulator of the protein C system. accelerated inhibition of a-thrombin, TT-thrombin, ac- Direct evidence for the involvement of protein C inhibitor is tivated protein C, factor Xa, urokinase, and chymo- lacking, as an inhibitor deficiency has yet to be documented. trypsin, but not plasma kallikrein. The abilityof gly- Protein C inhibitor is a member of the serine proteinase cosaminoglycans to accelerateproteinase inhibition inhibitor (serpin)’ superfamily of proteins, whose prototype appeared to depend on the formationof a ternary com- is a’-proteinase inhibitor (7). Protein C inhibitor can be plex of inhibitor, proteinase, and glycosaminoglycan. further classified as a heparin-binding serpin, along with the The optimum heparin concentration for maximal rate proteinaseinhibitors antithrombin (historically known as stimulation varied from 10 to 100 pg/ml and was re- antithrombin 111) and heparin cofactor (also called heparin lated to the apparent affinity of the proteinase for cofactor 11). Heparin and some other glycosaminoglycans act heparin. There was no obvious relationship between to increase the rate of proteinase inhibition by these three heparin affinity and maximum inhibition rate or de- plasma inhibitors, insome cases as much as several thousand- gree of rate enhancement. The affinityof the resultant fold ($). The mechansim whereby heparin catalyzes protein- protein C inhibitor-proteinase complex was also not ase inhibition is a subject of much study, especially dueto the related toinhibition rate enhancement, and the results widespread use of heparin as a therapeuticanticoagulant. The showed that decreased heparin affinityof the complex ability of heparin to accelerate the inhibition of activated is not ap important part of the catalytic mechanism of protein C by protein C inhibitor, therebyfavoring coagulation, heparin. The importance of protein C inhibitor as a regulator of the protein C system may depend on the is at odds with the anticoagulant effect of heparin therapy. relatively large increase in heparin-enhanced inhibi- As a first steptoward understanding this apparentcontradic- tion rate for activated protein C compared to other tion and in order to gather insight into the physiological proteinases. importance of protein C inhibitor, a series of studies was undertaken. The work presented here describes the heparin- binding site of protein C inhibitor, the polyanion specificity of protein C inhibitor, and the mechanism whereby heparin Hemostasis requires a balance between procoagulant and accelerates proteinase inhibition. Some of these results have anticoagulant forces. Among the anticoagulant mechanisms appeared previously in abstract form (9). The following report is the protein C system. Thrombin generated during coagu- compares protein C inhibitor to antithrombin and heparin lation binds to thrombomodulin on vesselwalls andthe cofactor (10). thrombin-thrombomodulin complex activates the zymogen protein C. Activated protein C, with its cofactor protein S, EXPERIMENTALPROCEDURES (1). proteolytically inactivates coagulant factors V and VI11 Materiab-Human protein C inhibitor was purified as previously The importance of the protein C system is demonstrated by described (ll), as were antithrombin and heparin cofactor (12). All the incidence of thrombosis in individuals who lack protein C proteinases wereof human origin, with the exception of bovine (2) or protein S (3). The protein C system is believed to be chymotrypsin. a-Thrombin was purified as described (13) or con- verted to ?,-thrombin (14). Protein C was prepared as a by-product * This work was supported in part by Research Grant HL-06350 of the factor X preparation (15), and further purified and activated from the National Institutes of Health. The costs of publication of by incubating it with thrombin, then passing the mixture over im- this article were defrayed in part by the payment of page charges. mobilized antithrombin to remove thrombin and traces of factor Xa. This article must therefore be hereby marked “advertisement” in Urokinase was purchased from Sigma, plasma kallikrein from Calbi- accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $TO whom correspondence and reprint requests should be ad- The abbreviations used are: serpin, serine proteinase inhibitor; dressedThe University of North Carolina, Div. of Hematology, HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonicacid; PEG, Campus Box 7035, Chapel Hill, NC 27599-7035. Fax: 919-966-7639. polyethylene glycol; HPLC, high performance liquid chromatography.

8789

This is an Open Access article under the CC BY license. 8790 Protein to Binding Heparin C Inhibitor ochem, bovine chymotrypsin from Cooper Biomedical, and neutrophil 5-ml column of heparin-Sepharose in 20 mM HEPES, 10 mM NaC1, elastase from Elastin Products (Pacific, MO). The following protein- 0.1% PEG, pH 7.4. Samples of protein C inhibitor, proteinase, and ase substrates were used Chromozym TH (tosyl-Gly-Pro-Arg-p-ni- protein C inhibitor-proteinase complexes were eluted with a 1 ml/ troanilide) for thrombin from Boehringer Mannheim, Spectrozyme min linear salt gradient from 10 mM to 1.2 M NaCl; 0.25-ml fractions PCa (Lys(Cbo)-Pro-Arg-p-nitroanilide)for activated protein C and were collected. Proteins were detected by absorbance at 280 nm and Spectrozyme FXa (MeO-CO-CHG-Gly-Arg-p-nitroanilide)for factor by reactivity with a rabbitpolyclonal antiserum to protein Cinhibitor Xa from American Diagnostica, S-2444 (Glu-Gly-Arg-p-nitroanilide) (to detect protein Cinhibitor and proteinC inhibitor-proteinase for urokinase and S-2302 (Pro-Phe-Arg-p-nitroanilide) for kallikrein complexes) and by chromogenic substrate hydrolysis (to detect pro- from KabiVitrum, and Suc-Ala-Ala-Pro-Phe-p-nitroanilidefor chy- teinases and protein C inhibitor-proteinase complexes; an extended motrypsin from Sigma. The following were purchased from Sigma: incubation time was required to detect proteinase activity in com- bovine serum albumin, polybrene (1,5-dimethyl-1,5-diazaundecame- plexes). Results were plotted and the saltconcentration correspond- thylene polymethobromide), bovine heparan sulfate, bovine chon- ing to peak elution was determined. The mean and standard deviation droitin sulfate A, shark chondroitin sulfate C, fucoidan (a sulfated were calculated from multiple runs (2-10) of each sample. Protein C polymer of fucose from a marine alga), phosvitin (a phosphoserine- inhibitor-proteinase complexes were prepared by incubating protein- containing glycoprotein from egg yolk), sulfatides, and tetrapoly- ase with a slight excess of protein C inhibitor for a time previously phosphate. Unfractionated heparin was from Diosynth (Oss, the determined to allow complete reaction, at least five times the half- Netherlands). Low molecular weight (M,5500) heparin was from life of the reaction. Inactivation of protein C inhibitor by neutrophil Calbiochem. Dermatan sulfate was purchased from Calbiochem and elastase was followedby the loss of thrombin inhibition activity. treated with nitrous acid to remove contaminatingheparin (16). DSPG I1 (a dermatan sulfate proteoglycan from bovine skin) was the RESULTS gift of Dr. Lawrence Rosenberg, Montefiore Hospital, Bronx, NY. Heparin fractions with low and high affinity for antithrombin were Identification of the Heparin-binding Site of Protein C In- the gift of Dr. Ingemar Bjork, Swedish University of Agricultural hibitor-Because heparin is a negatively charged glycosami- Sciences Uppsala, Sweden. Chemically depolymerized heparin (av- noglycan, it is expected to bind to basic residues of protein C erage molecular weight 5000, 3500, and 2500) was the gift of Dr. inhibitor. Three regions were identified as potential heparin- Charles Griffin, Miami University, OH. Fmoc-amino acids were from MilliGen. Heparin-Sepharose was from Pharmacia LKB Biotechnol- binding sites: residues 1-11, 82-90, and 266-278. Residues ogy Inc. and heparin-agarose was prepared as described (12). 266-278 were previously identified as a heparin-binding site Peptide Synthesis-Peptides were assembled from Fmoc amino by homology to a consensus glycosaminoglycan recognition acids using a Milligen pepsynthesizer according to thereported cDNA site (18). Peptide 1-11 also follows the consensus sequence. sequence of protein C inhibitor (17). In peptide 1-16, tyrosine was Sequences corresponding to the three regions of protein C added to the carboxyl terminus, and in peptide 80-93, Phe-92 was inhibitor (1-16,80-93, and 264-283, and the264-283 random substituted with tyrosine to facilitate spectrophotometric quantita- tion of the peptides. Purity of the peptides was verified by reverse- sequence) are shown in Fig. 1, along with helical-wheel pro- phase HPLC, and when necessary, further purification was accom- jections of these sequences (19). The projections show that plished by HPLC. Heparin affinity of peptides was measured using a the selected sequences can form amphipathic helices with 1-ml column of heparin-agarose in 20 mM HEPES, 0.1% PEG 8000, basic and uncharged residues on opposite faces of the helix. pH 7.4, with a linear salt gradient. To test whether the selected sequences represent the hep- Peptide CompetitionAssays-Peptides (0.1-100 FM) in 20mM arin-binding site of protein C inhibitor, the three peptides HEPES, 150 mM NaCl, 0.1% PEG, pH 7.4, were added to assay mixtures containing 1pg/ml heparin, 2 mg/ml bovine serum albumin, were synthesized and their ability to bind heparin was tested and 50 or 100 nM protein C inhibitor in the same buffer at 25 "C. in proteinase inhibition assays. The peptides were added to Reactions were started with the addition of 5 nM thrombin or 10 nM systems containing protein C inhibitor, proteinase, and hep- activated protein C. After incubation, remaining proteinase activity arin; with the premise that heparin-binding peptides would was measured using 0.15 mM chromogenic substrate with 2 mg/ml compete with protein C inhibitor for heparin, thus decreasing Polybrene. Substrate hydrolysis was linearly related to proteinase concentration. Control experiments verified that none of the peptides affected either the ability of proteinase to hydrolyze substrate or the ability of protein C inhibitor to inhibit proteinases in the absence of heparin. Polyanion Accelerationof Protein CInhibitor Activity-Polyanions were dissolved in 20 mM HEPES, 150 mM NaC1, 0.1% PEG, pH 7.4, and various concentrations (0.1-1000 pg/ml) were added to 60 nM protein C inhibitor and 2 mg/ml bovine serum albumin in the same buffer at 25 "C. Reactions were started with the addition of 5 nM PEPTIDE 1-18 PEPTIDE 80-93 thrombin or activated protein C, incubated for various times, and HRHHPREMKKRVEDLH LOKSSEKELHRQVO remaining proteinase activity was measured. Control experiments in the absence of protein C inhibitorverified that none of the polyanions directly affected the ability of proteinase to hydrolyze substrate. Second order proteinase inhibition rate constants were calculated as -In a/t [proteinC inhibitor] where a is the fractional proteinase activity remaining relative to theuninhibited control, t is the time of incubation, and [protein C inhibitor] is the inhibitor concentration, 60 nM. The inhibition rate increase was calculated by dividing the maximum inhibition rate in the presence of polyanion by the rate in the absence of polyanion, and the optimum polyanion concentration PEPTIDE 264-283 RANDOM PEPTIDE -283 was the concentration that produced the maximum inhibition rate. SEKTLRKWLKMFKKROLELV TRMKOLKLFKWERLEKLKSV Heparin Acceleration of Proteinase Inhibition-Heparin (0.1-1000 FIG. 1. Potential heparin-binding regions of protein C in- pg/ml) was added to 50 nM protein C inhibitor in 20 mM HEPES, hibitor. Three regions rich in basic residues were selected and 150 mM NaC1, 0.1% PEG, pH 7.4, containing 2 mg/ml bovine serum peptides corresponding to these regions were synthesized. The se- albumin at 25 "C. Reactions were started by adding 5 nM proteinase quences of the peptides are shown along with helical wheel diagrams (a-thrombin, yT-thrombin, activated protein C, factor Xa, urokinase, of the sequences (assuming 3.6 residues/helical turn). In peptide 1- plasma kallikrein, or chymotrypsin). Incubation times varied among 16, a tyrosine residue not found in native protein C inhibitor was proteinases and remaining proteinase activity was determined by added to thecarboxyl terminus. Inpeptide 80-93, Phe-92 was replaced substrate hydrolysis. Proteinase inhibition rate constants were cal- with tyrosine in the synthetic peptide. The random sequence peptide culated as described above. was constructed using the same residues as peptide 264-283, but in Heparin Affinity Chromatography-Experiments were performed random order. The first and last residues of peptide 264-283 and the using a Pharmacia fast protein liquid chromatography system and a random sequence peptide are notvisible in the helical wheel diagrams. H eparin Binding to Protein to Binding Heparin C Inhibitor 8791 the rate of the proteinase inhibition reaction. The results TABLEI revealed that of the three peptides, only peptide 264-283 Polyanwn acceleration of proteinuse inhibition by protein C inhibitor competed with protein C inhibitor for binding to heparin, as ThrombinActivated protein C shown by the increase in proteinase activity with increasingRate Optimum RateOptimum Polyanion peptide concentration (Fig. 2). This effect was observed for (polyanion)”increaseb (polyanion) increase both thrombin and activated protein C. A peptide comprised dm1 dm1 of the same residues as peptide 264-283, but in random Heparin 10 9.4 100 52.2 sequence, did not compete as effectively, suggesting that the Heparan sulfate 300100 1.9 3.1 three-dimensional distribution of charged residues rather Chondroitin sulfate A >lo00 1.7 >lo00 4.1 than the number of charged residues was responsible for its Chondroitin sulfate C ND‘ ND heparin binding function. As further confirmation of peptides Fucoidan 10 9.5 100 10.0 models for the heparin binding region in proteins, peptide Dermatan sulfate ND >lo00 4.6 as DSPG I1 ND ND 264-283 interfered with heparin-catalyzed inhibition of Phosvitin 30 1.6 100 5.4 thrombin by twoother heparin-binding inhibitors, antithrom- AT low affinity hepa- 30 3.0 100 10.3 bin and heparin cofactor, and peptides corresponding to the rin putativeantithrombin heparin-binding site (Ala-124-Leu- AT high affinity hepa- 30 3.5 100 8.7 140, Ref. 10 ) competed for heparin binding in the protein C rin Low molecular weight 10 2.6 30 12.1 inhibitor assay (resultsnot shown). Furthermore, peptide heparin 264-283 bound to immobilized heparin and was eluted with Heparin M, 5000 10.2 303 1.2 800 mM NaC1, indicating even higher affinity than native Heparin M, 3500 10 1.5 100 12.5 protein C inhibitor (elution at approximately 500 mM NaC1) Heparin M, 2500 ND 300 4.3 The random sequence peptide bound less tightly (elution at “The optimum polyanion concentration is the concentration at 380 mM NaC1, Ref 10). which the maximum proteinase inhibition rate occurs. Polyanwn Specificity of Protein C Inhibitor-In addition to * The rate increase is calculated as the ratio of the maximum rate heparin, some other glycosaminoglycans have been found to to therate in the absence of polyanion. accelerate proteinase inhibition by protein C inhibitor (20, e ND, no acceleration of inhibition detected at 1 mg/ml polyanion. 21). To further assess the specificity of protein C inhibitor, we tested the ability of various polyanions to catalyze protein vated protein C inhibition by protein C inhibitor, although C inhibitor inhibition of thrombin and activated protein C. the effect of chondroitin sulfate A was weak, as shown by the Each polyanion was tested at various concentrations in order high concentration required to increase the inhibition rate. to determine the maximum degree of rate enhancement and Dermatan sulfate showed no activity with thrombin and very the polyanion concentration at which that occurred. The low activity with activated protein C; a highly purified der- resultsare presented in Table I. Among the mammalian matan sulfate preparation (DSPG 11) showed no activity with glycosaminoglycans, onlyheparin, heparan sulfate, and chon- either proteinase. Chondroitin sulfate C did not accelerate droitin sulfate A caused acceleration of thrombin and acti- inhibition. Interestingly, fucoidan (a sulfated polymer of fu- cose from a marine alga) was relatively effective, and phos- vitin (a phosphoserine-containingglycoprotein from egg yolk) was also active. Sulfatides and tetrapolyphosphate did not affect the rate of proteinase inhibition by protein C inhibitor (not shown). A number of heparin fractions accelerated inhibition of thrombinand activated proteinC by protein Cinhibitor (Table I). Protein C inhibitorclearly differs from antithrom- bin in responding equally well to heparin with high or low affinity for antithrombin. Protein C inhibitorexhibited some 0.1 1 10 100 specificity for the size of the heparin molecule, as greater Ip.PUd.1 wnr) concentrations of the smaller heparin fractions were required for maximum stimulation of the inhibition reaction. Mechanism of Heparin Acceleration of Inhibition-The mechanism whereby heparin accelerates proteinase inhibition by protein C inhibitorcould depend on heparin binding to the inhibitor, to theproteinase, or to both proteins. Examination of the effect of increasing concentrations of heparin on the rate of inhibition of six different proteinases (a-thrombin, Y~- thrombin, activated protein C, factor Xa, urokinase, and chymotrypsin, Fig. 3) reveals a bell-shaped curve that is I consistent with a ternary complex model for heparin action 0.1 1 10 loo FWWW) (22-24). Accordingto thismodel, heparin binds both inhibitor FIG. 2. Peptide competition in proteinase inhibition assays. and proteinase to bring the reactants into closer proximity, Synthetic peptides corresponding to potential heparin-binding re- and the rate of reaction increases as heparin concentration gions in protein C inhibitor (shown in Fig. 1)were added to thrombin increases. When heparin concentrations increase beyond a (panel A) or activated protein C (panel B) inhibition assays as certain point, the proteinase and inhibitor are more likely to described under “Experimental Procedures.” They axes show throm- bind to separate heparin molecules, thus decreasing the cat- bin and protein C activity relative to the activity in the absence of inhibitor. Greater proteinase activity indicates diminished inhibition alytic efficiency of the glycosaminoglycan. If heparin binding due to peptides competing with proteins for heparin binding. B, to only one of the two reactants were important, then heparin peptide 1-16; 0,peptide 80-93; +, peptide 264-283; 0,random peptide would exhibit a saturationphenomenon in these experiments. 264-283. The optimum heparin concentration as well as the relative 8792 Heparin Binding toProtein C Inhibitor TABLEI1 Proteinase heparin affinity and inhibition rate increase Maximum Rate Proteinase Heparin Heparin affinity” optimumb rate increase‘

Thrombin 640 f 27 10 11.2 9.4 Factor Xa 4460.21 f 19 30 1.1 Activated protein C 357 f 7 100 1.54 52.2 Urokinase 344 f 7 100 0.02 2.3 Chymotrypsin 290f 28 100 203 3.0 1 10 loo a Heparin affinity is given as the salt concentration required for Heparin (pg/rnl) peak elution from immobilized heparin. This value is an indication of heparin affinity,not a true affinity constant. The optimum heparin concentrationis the concentration at which the maximum inhibition rate occurs. The rate increase is calculated as the ratio of the maximum rate to the rate in the absence of heparin.

TABLEI11 Heparin affinityof protein C inhibitor and protein C inhibitor- proteinase complexes I Inhibitor-proteinasecomplex Salt conc. at peak elution“ 1 10 100 mM NaCl Heparin (pg/rnl) Protein C inhibitorb 508 f 11 FIG. 3. Heparin acceleration of proteinase inhibition by Protein C inhibitor-thrombin 562 f 15 protein C inhibitor. The effect of heparin on the rate of inhibition Protein C inhibitor-activatedprotein C 448 f 25 was measured as described under “Experimental Procedures.”Panel Protein C inhibitor-factor Xa 525 f 17 A: W, a-thrombin; 0, 7,-thrombin; +, activated protein C. Panel B: Protein C inhibitor-urokinase 422 f 28 W, factor Xa; 0 urokinase; +, plasma kallikrein; 0, chymotrypsin. The salt concentration required for peak elution from immobilized The y axis shows the relative inhibition rate, the ratio of the rate heparin. constant at a particular heparin concentration to the rate constant ‘Protein C inhibitoraffinity was measured in the absence of in the absence of heparin. proteinase. increase in inhibition ratevaried among different proteinases. arin was most efficient at catalyzing the inhibitionof throm- All experiments were performed with identical concentrations bin, which hadthe highest heparinaffinity. There is no of protein C inhibitor and proteinase in order to rule out obvious relationship between heparin affinity and the maxi- protein concentration-dependent effects on heparin optimum mum rate of inhibition or the degree of rate enhancement. or maximum rate. Interestingly, the rateof inhibition of factor Kallikrein, which exhibited no inhibition rate enhancement, Xa was diminished at low heparin concentrations, and at did bind to immobilized heparin, eluting at 305 mM NaCl. optimum heparin the ratewas only slightly greater than the The salt concentration whichat a protein elutes from heparin rate in the absenceof glycosaminoglycan. In addition, increas- is an indication of heparin affinity and is useful for compar- ing concentrationsof heparin progressively decreasedthe rate ative purposes, but is not identical to an affinity constant, of inhibition of plasma kallikreinby protein C inhibitor (Fig. which requires more rigorous measurement. 3). These two cases suggestthat the catalyticeffect of heparin Table I11 describes the apparent heparin affinityof protein on proteinC inhibitor reactivity is nota simple phenomenon, C inhibitor following reactionwith four proteinases.The but might involve additional effects of heparin on the partic- reaction between protein C inhibitor and its target proteinases ular inhibitor-proteinase pair. results in a stable bimolecular complex (6). Complexes con- The failure of attempts to demonstrate the importance of taining protein C inhibitor and one of four proteinases were simultaneousbinding of both inhibitor and proteinase to applied to immobilized heparin.The protein C inhibitor- heparin using a kinetics approach (25, 26) was most likely due to the fact that heparin elicited a relatively mild rate thrombin complex eluted at higher salt concentration than increase. This meant that rate enhancementwas not detect- protein C inhibitor alone, most likely due to the contribution ableunder the required conditions, where proteinaseand of thrombin, whose affinity for heparin was even higher. The inhibitorconcentrations must saturate a small amount of protein C inhibitor-factor Xa complex eluted at essentially heparin. thesame salt concentration as protein C inhibitor alone, Heparin affinity chromatographywas used to further assessalthough the affinity of factor Xa itself was lower than that the contribution of heparin bindingto proteinC inhibitor and of protein C inhibitor. The othercomplexes, protein C inhib- various proteinases before and after the inhibition reaction. itor-activated protein C and protein C inhibitor-urokinase, Protein C inhibitor and five proteinases (a-thrombin, acti- showed heparin affinities intermediate to those of protein C vatedprotein C, factor Xa, urokinase, and chymotrypsin) inhibitor and the individual proteinase. It is obvious from bound toimmobilized heparin andwere eluted by salt concen- these results that there is no uniform decrease in apparent trations greater thanphysiological (290-640 mM NaC1, Table heparin affinityof protein C inhibitor following reaction with 11). Therefore, in inhibition assays (containing150 mM NaC1) proteinase. Nor does the change in heparin affinity correlate heparin binding would occur. Table I1 demonstrates the in- with the degree of inhibition rate enhancement caused by verse relationship between the apparent heparin affinityof a heparin (listed in Table 11). Protein C inhibitor inactivated particular proteinase and the heparin concentrationrequired with neutrophil elastase, presumably at or near the reactive for maximum inhibition rate enhancement;for example, hep- site, eluted from heparin-Sepharoseat 530 f 14 mM NaC1. H eparin Binding to Protein to Binding Heparin C Inhibitor 8793

DISCUSSION (33).*The chemical heterogeneity and difficult purification of glycosaminoglycans from different tissues and species could These studies were undertaken in order to better under- contribute to the discordant findings. stand how heparin binds to protein C inhibitor and how this The degree of acceleration of activated protein C or throm- is important for the physiological function of protein C inhib- bin inhibition by protein C inhibitor varied with the polyan- itor. The results focus on three areas: identification of the ion. In any case, the increase in rate is relatively mild when putative heparin-binding site of protein C inhibitor, the gly- protein C inhibitor is compared to antithrombin or heparin cosaminoglycan specificity of protein C inhibitor, and the mechanism whereby heparin accelerates proteinase inhibi- cofactor (IO), although the increase in rate appears to depend on the particular reaction conditions, as different results have tion. A region of the protein C inhibitor molecule, residues been reported for the heparin enhancement of protein 264-283, was identified asthe heparin-binding site. This C inhibitor activity (11, 34). The experiments in the present sequence can be represented as an amphipathic helix, with study were performed with constant concentrations of pro- basic and uncharged residues on opposite faces of the helix, which is a property of other heparin-binding peptides (18). A teins in order to facilitate comparisons among polyanions. synthetic peptide corresponding to this region of protein C The concentration of a polyanion required for maximum rate inhibitor bound to heparin and interfered in heparin-cata- enhancement also varied; this most likely reflects the affinity inhibitor and the lyzed proteinase inhibition. This sequence, located in the H of the glycosaminoglycan for protein C helix of protein C inhibitor (7), clearly differs from the hep- proteinases. Among the polyanions tested, the optimum con- centrations varied in parallel for thrombinand activated arin-binding sites of the related proteins antithrombin and protein heparin cofactor, which have been assigned primarily to the C inhibition; this suggests a specific interaction of the polyanions with the inhibitor. D helix on the basis of chemical modification experiments The effect of various concentrations of heparin on the rate and natural mutations (7). The present results also differ of inhibition of five proteinases (thrombin, factor Xa, acti- from a recent report implicating the amino terminus of pro- vated protein C, urokinase, and chymotrypsin) by protein C tein C inhibitor (theA+ helix) in heparin binding (27). Failure inhibitor was consistent with the ternary complex model (22) to detect heparin binding of synthetic peptide 1-16 in the in which both proteinase and inhibitor bind to the glycosa- present experiments could be due to the assumption that the minoglycan. This phenomenon has been observed previously protein C inhibitor heparin-binding site is simply a linear for protein C inhibitor (20, 21, 34) and is well-characterized sequence of residues and thatsynthetic peptides would adopt for the related inhibitors antithrombin and heparin cofactor. the same structure (and therefore function) as the sequences Most of the polyanions that stimulated proteinase inhibition in the native protein. It is also possible that the monoclonal in the present study and in other reports appear to follow the antibody used in the previous study, which bound to helix same mechanism. In no case was a saturation effect observed, A+, might interfere with heparin binding to theH helix; this although it is difficult to completely rule out this possibility is certainly a possibility if the and the H helices together A+ in cases where stimulation is weak and polyanion concentra- form a heparin-binding site, as was suggested (27). tions cannot be increased further. Acceleration of protein C inhibitor proteinaseinhibition by This work represents a further attempt to correlate the a variety of glycosaminoglycans (especially heparin, heparan mechanism of heparin action with heparin binding properties sulfate, fucoidan, and low molecular weight heparin) and other of the proteins involved. The heparin concentration required polyanions (phosvitin) is consistent with a relatively nonspe- for maximum inhibition rate enhancement is a function of cific heparin-binding site in protein C inhibitor. This is in the apparent affinity of the proteinase for heparin, but there contrast to antithrombin, which exhibits narrow specificity appears to be no general rule governing the magnitude of the for heparin and heparansulfate (28). Protein C inhibitor more maximum rate or the rate increase. The rate of inhibition of closely resembles heparin cofactor, which allows an even wider factor Xa was actually decreased at low concentrations of variety of polyanions to accelerate thrombin inhibition (29- heparin, and the rate of inhibition of plasma kallikrein pro- 31). Previous studies noted the ability of several sulfated gressively decreased with increasing heparin. One explanation glycosaminoglycans to enhance protein C inhibitor activity, for these results is that in the case of thrombin, activated although no effect of heparan sulfate was detected (20, 21). protein C, urokinase, or chymotrypsin, the primary role of The ability of fucoidan (which contains no carboxyl groups) heparin is to bind both inhibitor andproteinase to bring them and phosvitin (which contains no sulfate groups) to enhance into close proximity formore rapid reaction than in the proteinase inhibition by protein C inhibitor is further evi- absence of heparin. However, in the case of factor Xa and dence for a relatively nonspecific glycosaminoglycan-binding kallikrein, other factors such as ionic interactions, confor- site in protein C inhibitor. Interestingly, the “permissive” mational changes, or steric hindrance might partially or com- glycosaminoglycan-binding site of protein C inhibitor did not pletely offset the rate enhancement due to the proximity accommodate dermatan sulfate well; this glycosaminoglycan effect. Different heparin-enhanced rate constants have been is the most effective at increasing the inhibition rate of reported previously (6,11, 34); some of the differences are heparin cofactor (32). The identity of the glycosaminoglycan most likely due to reaction conditions, particularly tempera- that accelerates thrombin or activated protein C inhibition tureand protein concentration, as well asthe source of by protein C inhibitor in vivo remains a mystery, although heparin. The concentrations of inhibitor and proteinase have heparan sulfate, which lines vessel walls and is believed to be been shown to influence the optimum heparin concentration the primary antithrombin-activating glycosaminoglycan (28), and maximum rate enhancement (26). In the present study is a candidate. Protein C inhibitor, however, did not distin- experiments with different proteinases were performed with guish heparin of high or low affinity for antithrombin. The the same protein concentrations to eliminate other variables. low activity of chondroitin sulfate A measured in this study It has been reported that theinhibition of activated protein could be due to heparan sulfate contamination, asthis glycos- aminoglycan was not treated with nitrous acid to destroy We found that dermatan sulfate that had not been treated with traces of heparin (16). The present study is at odds with a nitrous acid to destroy contaminating heparin did accelerate uroki- recent report that dermatan sulfate from various sources nase inhibition by protein C inhibitor, but acid-treated dermatan accelerated inhibition of urokinase by protein C inhibitor sulfate had no effect on the urokinase inhibition rate. 8794 BindingHeparin to Protein C Inhibitor

C by aprotinin is enhanced by heparin (35); this suggests that and Bertina, R. M. (1987) Ann. Znt. Med. 106,677-682 heparin might directly affect the catalytic properties of acti- 4. Heeb, M. J., Espana, F., Geiger, M., Collen, D., Stump, D. C., vated protein C. Inhibition of activated protein C by protein and Griffin, J. H. (1987) J. Biol. Chem. 262, 15813-15816 5. Heeb, M. J., Gruber A., and Griffin, J. H. (1988) J. Biol. Chem. C inhibitor is accelerated by heparin to a much greater degree 266, 17606-17612 than for any other proteinase tested, although no effect of 6. Suzuki, K., Nishioka, J.,Kusumoto, H., and Hashimoto, S. (1984) heparin on hydrolysis of a small peptide substrate by activated J. Biochem. (Tokyo) 95, 187-195 protein C was detected in the present study. 7. Huber, R., and Carrell, R. W. (1989) Biochemistry 28,8951-8966 It has been proposed that part of the ability of heparin to 8. Bjork, I., and Danielsson, A. (1986) in Proteinase Inhibitors (Barrett, A. J., and Salvesen, G. S., eds) pp. 489-513, Elsevier, dramatically accelerate thrombin inhibition by antithrombin Amsterdam is due tothe decreased heparin affinity of the resultant 9. Pratt, C. W., and Church,F. C. (1991) FASEB J. 5,1540 (abstr.) antithrombin-thrombin inhibitory complex, which allows 10. Pratt, C. W., Whinna, H. C., and Church, F. C. (1992) J. Biol. heparin to efficiently dissociate from the proteins in order to Chem. 267,8795-8801 participate in additional rounds of catalysis (36). (Native 11. Pratt, C. W., Macik, B. G., and Church, F. C. (1989) Thromb. antithrombin eluted from heparin-Sepharose at 925 mM Res. 53,595- 602 12. Griffith, M. J., Noyes, C. M., and Church, F. C. (1985) J. Bid. NaC1, while the antithrombin-thrombin complex and anti- Chem 260,2218-2225 thrombin inactivated at the reactive site by neutrophil elas- 13. Church, F. C., and Whinna, H. C. (1986) Anal. Biochem. 157, tase eluted at 430 mM NaCl, in agreement with a previous 77-83 report (36); results not shown.) This phenomenon is probably 14. Stone, S. R., Braun, P. J., and Hofsteenge, J. (1987) Biochemistry less important for protein C inhibitor inhibition of protein- 26,4617-4624 15. Griffith, M. J., Breitkreutz, L., Trapp, H., Briet, E., Noyes, C. ases, asthere wasno consistent correlation of maximum M., Lundblad, R. L., and Roberts, H. R. (1985) J. Clin. Znuest. inhibition rate or rate enhancementwith changes in apparent 75,4-10 heparin affinity of different protein C inhibitor-proteinase 16. Teien, A. N., Abildgaard, U., and Hook, M. (1976) Thromb. Res. complexes. The absence of a change in heparin affinity of the 8, 859- 867 protein C inhibitor-activated protein C complex has been 17. Suzuki, K., Deyashiki, Y., Nishioka, J., Kurachi, K., Akira, A., Yamamoto, S., and Hashimoto, S. (1987) J. Biol. Chem 262, previously noted (37). 611-615 In conclusion, protein C inhibitor clearly belongs to the 18. Cardin, A. D., and Weintraub, H. J. R. (1989) Arteriosclerosis 9, group of serpins whose reactivity toward certain proteinases 21-32 is stimulated by heparin. (A comparison of three of these 19. Schiffer, M., and Edmunson, A. B. (1967) Biophys. J. 7, 121-135 proteins is contained in the following paper, Ref. 10.) The 20. Kazama, Y.,Niwa, M., Yamagishi, R., Takahashi, K., Sakura- mechanism of heparin acceleration of proteinase inhibition gawa, N., and Koide, T. (1987) Thromb. Res. 48, 179-185 21. Kazama, Y., Koide, T., and Sakuragawa, N. (1989) Thromb. Res. by protein C inhibitor was consistent with a ternary complex 54,499-504 model, but there were indications that the exact mechansim 22. Griffith, M. J. (1982) J. Biol. Chem 257, 7360-7365 is a more complicated function of the particular proteinase 23. Peterson, C. B., and Blackburn,M. N. (1987) J. Biol. Chem 262, involved, as shown by the large rate enhancement for acti- 7559- 7566 vated protein C inhibition relative to other proteinases. The 24. Olson, S. T. (1988) J. Biol. Chem 263, 1698-1708 25. Griffith, M. J. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 5460- question remains how protein C inhibitor can regulate the 5464 anticoagulant protein C system when it is also an inhibitor of 26. Nesheim, M. E. (1983) J. Biol. Chem 258, 14708-14717 procoagulant enzymes. So far there is little evidence for a 27. Kuhn, L. A., Griffin, J. H., Fisher, C. L., Greengard, J. S., Bouma, protein C system-specific glycosaminoglycan that might pref- B. N., Espana, F., and Tainer, J. A. (1990) Proc. Natl. Acad. erentially accelerate activated protein C inhibition, since all Sci. U. S. A. 87,8506-8510 of the polyanions tested in the present study stimulate inhi- 28. Bauer, K. A,, and Rosenberg, R. D. (1991) Semin. Hematol. 28, 10-18 bition of both thrombin and activated protein C. It is also 29. Church, F. C., Treanor, R. E., Sherrill, G. B., and Whinna, H. C. possible that additional factors, such as the relative concen- (1987) Biochem. Biophys. Res. Commun. 148, 362-368 trations of activated protein C, thrombin, and otherenzymes 30. Church, F. C., Pratt, C. W., Treanor, R. E., and Whinna, H. C. or their localization in. uiuo, are critical for determining the (1988) FEBS Lett. 237, 26-30 physiological effectiveness of protein C inhibitor. 31. Church, F. C., Meade, J. B., Treanor, R. E., and Whinna, H. C. (1989) J. Bid. Chem 264, 3618-3623 32. Tollefsen, D. M., Pestka, C. A., and Monafo, W. J. (1983) J. Biol. Acknowledgments-We thank Joan Woods, Herbert Whinna, and Chem 258,6713-6716 Dr. Tu1 Kalayanamit for technicalassistance, and Drs.Ingemar 33. Geiger, M., Priglinger, U., Griffin, J. H., and Binder, B. R. (1991) Bjork, Charles Griffin, and Lawrence Rosenberg for gifts of glycosa- J. Biol. Chem 266,11851-11857 minoglycans. 34. Espana, F., Berrettini, M., and Griffin, J. H. (1989) Thromb. Res. 55, 369- 384 REFERENCES 35. Espana, F., Estelles, A,, Griffin, J. H., Aznar, J., and Gilabert, J. (1989) Thromb. Res. 56, 751-756 1. Esmon, C. T. (1989) J. Biol. Chem. 264,4743-4746 36. Carlstrom, A.-S., Lieden, K., and Bjork, I. (1977) Thromb. Res. 2. Griffin, J. H., Evatt, B., Zimmerman, T. S., and Kleiss, A. J. 11, 785- 797 (1981) J. Clin. Inuest. 68, 1370-1373 37. Laurell, M., and Stenflo,J. (1989) Thromb. Haemostusis 62,885- 3. Engesson, L., Broekmans, A. W., Briet, E., Brommer, E. J. P., 891