Zinc as a cofactor for heparin neutralization by histidine-rich glycoprotein.

We have studied the ability of histidine-rich glycoprotein (HRG) to neutralize the anticoagulant activity of heparin in plasma and in a purified component clotting assay. Addition of HRG to plasma or to the purified component assay did not neutralize the anticoagulant activity of heparin unless micromolar concentrations of zinc were present. Higher zinc concentrations were required for citrated than for heparinized plasmas due to competition of citrate with HRG for zinc binding. Zinc concentrations as low as 1.25 microM revealed HRG to be a powerful competitor of antithrombin for heparin in the purified component assays. HRG binding of heparin also was shown by affinity chromatography of HRG from immobilized heparin in the presence and absence of zinc. In the absence of zinc, HRG was eluted by 0.1 M NaCl, but, in the presence of zinc, elution of HRG required 1.0 M NaCl. Investigation of other divalent cations (copper and magnesium) indicated that augmentation of heparin binding by HRG in the presence of antithrombin was restricted to zinc. The HRG.Zn complex effectively competes with antithrombin for heparin, which restricts the availability of heparin to bind antithrombin and allows thrombin-mediated fibrinogenesis to proceed unimpeded. This could be initiated by zinc released from activated platelets.

Our research has focused on the reported heparin-neutraliz-ing function of HRG and its potential importance in hemostasis by studying the effects of thrombin, HRG, antithrombin, heparin, and divalent cations on fibrin generation in plasma and in a purified component clotting assay. Our initial experiments with plasma confirmed previous findings that micromolar concentrations of zinc significantly increased heparin binding by HRG (4), and competition experiments performed by using the purified component clotting assay revealed that heparin binding by HRG in the presence of zinc was greater than heparin binding by antithrombin. This research has shown that zinc concentrations in normal plasmas are not sufficient to serve a cofactor function for HRG, thus allowing heparin to serve its cofactor function for antithrombin. However, micromolar concentrations of zinc, compatible with concentrations released from activated platelets, are sufficient to serve a cofactor function for HRG to bind heparin more effectively than antithrombin and thereby promote fibrinogenesis. These data prompt us to propose that this paradigm could function to control the locality and amount of fibrin deposited in tissues.

EXPERIMENTAL PROCEDURES
Proteins and Reagents-Human HRG was purchased from Celsus Laboratories (Cincinnati, OH) as a 100 g/ml solution with a mean molecular mass of 60 kDa. We also prepared HRG from fresh human plasma (see below), and this preparation is referred to as HRGk75. Human antithrombin (AT) was obtained from Celsus Laboratories (58 kDa, 150 g/ml), and from Enzyme Research Laboratories, South Bend, IN (58 kDa, 1.59 mg/ml). Bovine thrombin was purchased from Parke-Davis (Morris Plains, NJ) as USP Thrombostat (1000 units/ml), and human thrombin (1.05 mg/ml) was purchased from Enzyme Research Laboratories. Heparin (12-15 kDa, 10,000 USP units/ml) was obtained from Elkins-Sinn (Cherry Hill, NJ). Aprotinin was purchased from ICN Biomedicals (Aurora, OH). Phosphocellulose (P11) was purchased from Whatman International (Fairfield, NJ). Mouse IgG1 monoclonal antibody (MO35) to an epitope in the first 229 amino acids of the N-terminal region of HRG was purchased from PanVera Corporation (Madison, WI), and monoclonal antibody to high molecular weight kininogen was from Dr. J. A. McIntyre of this center. Glycerol was purchased from Fisher Scientific (Pittsburgh, PA). Human fibrinogen (type I), epichlorohyrin-activated heparin-agarose resin, EDTA, DTT, 6-amino-n-hexanoic acid, benzamidine hydrochloride, and ultra pure grades of zinc acetate and the chloride salts of zinc, copper, and magnesium were purchased from Sigma.
Preparation of HRG from Fresh Human Plasma-Human HRG was prepared according to the method of Rylatt et al. (20). Briefly, phosphocellulose P11 was prewashed in 1 N HCl and 1 N NaOH, and prepared in a column (2.5 ϫ 20 cm) in equilibration buffer (pH 6.8) containing 10 mM sodium phosphate, 10 mM EDTA, 10 mM DTT, 1 mM 6-amino-nhexanoic acid, 1 mM benzamidine hydrochloride, and 0.5 M NaCl. Blood from two normal adult male donors was collected into Vacutainer tubes containing acid-citrate-dextrose (71 mM citric acid, 85 mM sodium citrate, and 111 mM dextrose; Becton Dickinson, Rutherford, NJ) and immediately centrifuged at 1000 ϫ g for 15 min. The pooled plasmas (40 ml) containing 10 mM EDTA, 10 mM DTT, 1 mM phenylmethysulfonyl fluoride, 1 mM benzamidine hydrochloride, 1 mM 6-amino-n-hexanoic acid, 50 units/ml aprotinin, and 0.5 M NaCl was applied to the column. The column then was washed with 10 bed volumes of washing buffer containing 0.8 M NaCl in equilibration buffer, and HRG was eluted with 5 bed volumes of elution buffer containing 2.0 M NaCl in equilibration buffer at a flow rate of 1 ml/min. The eluted HRG was concentrated, equilibrated with Tris-buffered saline (TBS; 0.05 M Tris-HCl, 0.15 M NaCl, pH 7.4), aliquoted (100 g/ml), and lyophilized. The protein composition and purity of HRG were analyzed by electrophoresis in SDS-polyacrylamide (8%) gels (21), by silver staining of the gels (22), and by Western blotting (23) with anti-HRG monoclonal antibody. Alkaline phosphatase-conjugated goat anti-mouse IgG antibody (Sigma) was used to detect reaction of anti-HRG antibody to HRG.
Plasma and Serum Collection-Blood samples (20 -40 ml) for thrombin time measurements were drawn from six healthy males (23 to 44 years) and six healthy females (25 to 31 years) into 20-ml plastic syringes containing 2 ml of 0.1 M sodium citrate, or 10, 20, and 40 units of heparin, or no anticoagulant. Final concentrations thus were 0.01 M sodium citrate or 0.5, 1.0, or 2.0 units/ml heparin. The bloods were centrifuged (15 min, 1000 ϫ g), plasmas or sera were pooled according to the type and amount of anticoagulant, and each pool was aliquoted into 0.5-ml samples and stored at Ϫ80°C. Monthly determinations of HRG in these pools by single radial immunodiffusion (24) revealed a mean concentration of 100 Ϯ 10 g/ml, which is consistent with reported values in human plasma (1). All experiments were performed on fresh aliquots from these pools, so repeat freezing/thawing was avoided.
Measurement of Zinc by Atomic Absorption Spectrophotometry-Zinc in plasma and serum was measured according to the method of Butrimovitz and Purdy (25) by using an atomic absorption spectrophotometer equipped with a single element hollow cathode lamp and autosampler (Perkin-Elmer 5100 PC and AS-51). A computer interface was used to measure zinc at 213.9 nm with an air-acetylene flame. Plasma and serum samples were obtained as described above and diluted 1:5 with 1% (v/v) HCl. Zinc standard solutions were prepared with 1% (v/v) HCl containing 5% (v/v) glycerol.
Thrombin Thrombin Time Measurements of Plasmas Supplemented with HRG-These experiments were performed in triplicate in disposable Coa Screener cuvettes to which were added 50 l of plasma and 100 l of TBS, as above. The plasmas used were from the citrated and the heparinized plasma pools, and the amount of HRG added to each aliquot of plasma in the cuvettes was 0.0, 0.13 (2.5% increase), 0.25 (5% increase), 0.50 (10% increase), 1.0 (20% increase), or 2.0 g (40% increase). All cuvettes contained 150 l of solution consisting of 50 l of plasma and the above defined amounts of HRG in 100 l of TBS. The cuvettes were incubated 2 min at 37°C in the Coa Screener, and thrombin times were determined by adding 50 l of thrombin (0.18 units in TBS) and recording the time required for formation of a fibrin clot.
Thrombin Time Measurements of Plasmas Supplemented with Zinc-These experiments were performed in triplicate in disposable Coa Screener cuvettes to which were added 50 l of plasma and 100 l of TBS, as above. The plasmas used were from the citrated pool and aliquots from the citrated pool that were heparinized with 0.018 units of heparin/ml. Each cuvette was supplemented with 6.2, 12.5, 25, 50, 100, or 200 M zinc in TBS. Initial experiments used both zinc chloride and zinc acetate but no differences were found, so all zinc experiments employed zinc acetate unless stated otherwise. All cuvettes contained 150 l of solution consisting of 50 l of either citrated or citrated/ heparinized plasma and the above amounts of zinc in TBS. The cuvettes were incubated 2 min at 37°C in the Coa Screener, and thrombin times were determined by adding 50 l of thrombin (0.18 units in TBS) and recording the time required for formation of a fibrin clot.
Thrombin Time Measurements of Plasmas Supplemented with HRG and Zinc-These experiments utilized aliquots from the citrated plasma pool supplemented with a constant amount of zinc and increasing amounts of HRG, and each preparation was studied for its ability to neutralize the anticoagulant activity of 0.098, 0.108, and 0.130 units/ml heparin by using thrombin times. The experiments were performed in triplicate in disposable Coa Screener cuvettes. Each cuvette contained 150 l of solution consisting of 50 l of citrated plasma containing 150 M zinc and 0.0, 0.25, or 2.0 g of HRG and the above amounts of heparin in TBS. The cuvettes were incubated for 2 min at 37°C in the Coa Screener, and thrombin times were determined by adding 50 l of thrombin (0.18 units in TBS) and recording the time required for formation of a fibrin clot.
Measurement of the Effect of Citrate on the Availability of Zinc-Thrombin times were performed as above to measure the effect of citrate on the availability of zinc. The experiments utilized aliquots of heparinized (0.125 units/ml) and heparinized/citrated (2.5 mM sodium citrate) plasmas containing 0.0 -800 M supplemental zinc. Plots of the thrombin time results for heparinized and heparinized/citrated plasmas were used to display the effect of citrate on the availability of zinc. A second set of thrombin time experiments was performed with aliquots of heparinized (0.25 units/ml) plasma containing 0.0 -800 M zinc, as well as with heparinized (0.25 units/ml) plasma containing a constant amount (100 M) zinc and a range of citrate concentrations (0.0 -0.8 mM) to determine if supplemental citrate would bind and limit the availability of zinc as measured by a corresponding increase in thrombin times.
Purified Component Clotting Assay: Competition of HRG and AT for Heparin-These experiments utilized purified components that are involved in the thrombin-mediated conversion of fibrinogen to fibrin, and thrombin times were used to measure the rates of this conversion reaction as an index of the availability of heparin to serve as cofactor for AT to inhibit thrombin. All experiments were performed in triplicate in disposable Coa Screener cuvettes to which were added 50 l of thrombin (1.05 g in TBS) and 100 l of TBS containing one of the following: TBS only; AT (3.2 g); AT (3.2 g) plus heparin (0.02 units); AT (3.2 g) plus heparin (0.02 units) plus HRG (2.0 g).
Zinc concentrations in the cuvettes for each of the above four solutions were 0.0, 0.63, 1.25, 2.5, or 5.0 M. The cuvettes were incubated 2 min at 37°C in the Coa Screener, and thrombin times were determined by adding 50 l of fibrinogen (125 g in TBS) and recording the time required for formation of a fibrin clot. The thrombin and the AT used in these experiments were from Enzyme Research, and the HRG was HRGk75 shown by Western blotting and silver staining of polyacrylamide gels to consist predominantly of native (i.e. 75 kDa) molecule.
Effect of Zinc on Binding of HRG to Immobilized Heparin-Heparin coupled to epichlorohydrin-activated 4% beaded agarose (Sigma, H0402) was loaded into a 0.9 ϫ 5.0-cm glass column and washed with 10 bed volumes of 0.02 M Tris-HCl, 3.0 M NaCl buffer (pH 7.2) containing 1.0 mM EDTA and 0.05% NaN 3 . The column was equilibrated with 10 bed volumes of 0.02 M Tris-HCl, 0.01 M NaCl buffer (pH 7.2) containing 0.1 mM zinc (equilibration buffer), was loaded with 0.5 mg HRGk75 in 1.0 ml of equilibration buffer, and equilibrated with equilibration buffer at a flow rate of 0.5 ml/min. The effect of zinc on the binding of HRG to immobilized heparin was studied by determining how much HRG was removed from the heparin-agarose by 0.1, 1.0, and 3.0 M NaCl in the presence of zinc or the zinc chelator, EDTA. Elution profiles of HRG from the immobilized heparin matrix were studied by using the two following methods. Method A consisted of 0.1 M NaCl, 0. Comparison of Zinc, Copper, and Magnesium for Heparin Neutralization in the Purified Component Clotting Assay-The purified component clotting assay was performed as described above. In these experiments, the Coa Screener was used to measure the rates of thrombinmediated conversion of fibrinogen to fibrin as an index of the availability of heparin to serve its cofactor function for AT to inhibit thrombin, and Coa Screener cuvettes were used for the determination of thrombin times. All cuvettes contained 150 l of TBS solution containing thrombin (0.16 units), AT (2.8 g; Celsus), heparin (0.008 units), HRG (2.0 g; Celsus), and either none or 5.0 M zinc, copper, or magnesium as the chloride salts. For the control thrombin time measurements, experiments also were done in the absence of heparin. The cuvettes were incubated for 2 min at 37°C in the Coa Screener, and thrombin times were determined by adding 50 l of fibrinogen (125 g in TBS) and recording the time required for formation of a fibrin clot. The effects of zinc, copper, and magnesium on heparin neutralization were compared.

RESULTS
Biochemical Characterizations-SDS-polyacrylamide gel electrophoresis and silver staining of Parke-Davis thrombin revealed a major band of 37 kDa and several minor bands at Ն 50 kDa, and thrombin from Enzyme Research produced one major band of 37 kDa (data not shown). These thrombins were identical in their functional capacities to convert fibrinogen to fibrin. SDS-polyacrylamide gel electrophoresis and silver staining of the fibrinogen showed three major bands between 50 and 60 kDa, corresponding to the ␣, ␤, and ␥ chains (26), and AT from Celsus and Enzyme Research produced one major 58-kDa band (data not shown). The above proteins were found to be free of HRG by Western blotting with the mouse monoclonal antibody to HRG (data not shown). Silver staining of SDSpolyacrylamide gels of HRGk75 revealed mostly 75-kDa protein and small amounts of 50-kDa material, while analysis of Celsus HRG revealed much less 75-kDa material and bands at 67, 50, 48, and 40 kDa (Fig. 1A), which are consistent with earlier biochemical studies of this molecule (27). Western blotting with monoclonal antibody to HRG revealed reactivity with the 75-, 67-, 50-, 48-, and 40-kDa bands (Fig. 1B), and Western blots of our normal sera and plasma pools reacted with monoclonal antibody to HRG in almost identical qualitative and quantitative patterns as with HRGk75. In addition, Western blots of HRG75k, Celsus HRG, and AT did not react (data not shown) with monoclonal antibody to high molecular weight kininogen, which is another histidine-rich glycoprotein (28).
In summary, thrombin from Enzyme Research was electrophoretically more homogenous than thrombin from Parke-Davis, but they functionally were equivalent; AT from Enzyme Research and Celsus were electrophoretically equivalent, and Western blotting revealed that neither thrombin nor AT contained HRG and that neither thrombin nor AT nor HRG contained high molecular weight kininogen. The HRG from Celsus was less homogeneous and contained more degradation products than HRGk75, yet monoclonal antibody to HRG revealed that the fragments were of HRG origin, and our normal human plasmas were found to have the same distribution of fragments as those identified in HRGk75.
The Effect of HRG on Heparin Activity-Earlier investigators reported that HRG-depleted citrated plasmas less effectively resisted the anticoagulant effects of heparin (2, 4, 5, 29). After confirming these observations, we studied the role of HRG by supplementing rather than depleting plasmas of HRG. The plasmas used in these experiments were from the citrated and heparinized plasma pools, and the results of these experiments revealed no significant differences in thrombin times when plasma concentrations of HRG were increased from base line (i.e. 100 g/ml) to 40% increase of base line. Each supplemented plasma was tested separately six times, and the results consistently failed to show any correlation between increasing plasma HRG concentrations and changes in the anticoagulant effects of heparin (data not shown).
Measurement of Zinc by Atomic Absorption Spectrophotometry-Measurements of zinc concentrations were performed on all reagents, plasmas and sera, and on solutions used in the thrombin time assays. The results showed negligible amounts of zinc (Ͻ0.01 M) in the TBS, thrombin, and heparin. Undiluted plasma zinc concentrations were 8.6 Ϯ 1.8 M, and serum zinc concentrations were 11.6 Ϯ 1.6 M.
The Effect of Zinc on Heparin Activity-The interaction between heparin and HRG has been reported to require divalent cations (2,5,27). We studied the differential effects of zinc on thrombin-mediated fibrinogenesis in citrated and citrated/heparinized plasmas and found that the thrombin times of citrated plasmas were not changed by supplemental zinc but that the thrombin times of citrated/heparinized plasmas were shortened in a dose-dependent manner (Fig. 2). The zinc-induced reduction of the anticoagulant activity of heparin was found to be reversible, as demonstrated by the finding that the addition of heparin to citrated/heparinized plasmas containing Ն 25 M zinc increased the thrombin time yet again (Fig. 2, insert). In addition to illustrating a central role for zinc-induced modulation of HRG-heparin interactions, these results suggested that the zinc-catalyzed HRG binding of heparin was saturable, for titrations of heparin at 100 and 200 M zinc produced identical thrombin times (Fig. 2, insert).
The Effect of HRG and Zinc on the Availability of Heparin-Citrated plasmas supplemented with a constant amount of zinc (150 M) in Coa Screener cuvettes revealed an enhanced ability to shorten thrombin times when increased amounts of HRG were added to the cuvettes, even in the presence of increased amounts of heparin (Fig. 3), indicating a correlation between plasma HRG concentrations and heparin neutralization in the presence of micromolar concentrations of zinc. Indeed, a similar correlation was observed when much smaller concentrations of zinc were used (i.e. 12.5, 25, and 50 M), but this correlation was not observed when supplementary zinc was not added to the cuvette (data not shown). In summary, incremental decreases in heparinized thrombin times were produced by the addition of HRG and zinc, but were not produced by the addition of identical amounts of HRG in the absence of zinc.
The Effect of Citrate on the Availability of Zinc-To determine the effect of citrate on the availability of zinc, thrombin times were determined for heparinized (0.125 units/ml) and heparinized/citrated (2.5 mM) plasmas containing 0.0 -800 M zinc. Neutralization of the catalytic effect of heparin on AT in heparinized/citrated plasmas was found to require 400 -800 M zinc, while heparinized plasmas required only 100 M zinc (Fig.  4), indicating that citrate interacts with and limits the availability of zinc. Similarly, the thrombin time of heparinized plasma containing 0.25 units of heparin/ml was found to be maximally depressed by 100 M zinc (Fig. 5A) while incremental additions of citrate (0.0 -0.8 mM) to the heparinized plasmas reversed the effects of zinc and prolonged the thrombin times (Fig. 5B), suggesting that citrate bound the zinc and disallowed its cofactor effect on HRG to bind and sequester heparin from AT.
Purified Component Clotting Assay: Competition of HRG and AT for Heparin-This assay was performed in the Coa Screener and measured the rate of thrombin-mediated conversion of fibrinogen to fibrin in the presence and absence of AT, heparin, and HRG in concentrations of zinc that ranged from 0.0 to 5.0 M. This assay was designed to study the possibility of zincactivated HRG binding of heparin in a functional system that included AT and excluded other heparin-binding proteins. The results showed that neither zinc nor AT affected the thrombin-mediated conversion of fibrinogen to fibrin (E and Ⅺ in Fig. 6), but this reaction was prolonged significantly when heparin was added (Ç in Fig. 6) due to the well established cofactor function of heparin for AT in the inhibition of thrombin-mediated conversion of fibrinogen to fibrin. Note a small increase in thrombin time by AT-heparin in the presence of 0.63 M zinc, but no subsequent increases from 0.63 to 5.0 M zinc. In striking contrast, when HRG was added, the rate of thrombin-mediated conversion of fibrinogen to fibrin was shortened progressively and returned to control values at Ն 1.25 M zinc (q in Fig. 6), again indicating that zinc served as cofactor for the neutralization of heparin by HRG. It should be pointed out that in this purified component clotting assay, the molar ratio of HRG to AT in the cuvette (i.e. 1.0:2.2) was the same as the ratio of HRG to AT in normal plasma (i.e. 1.0:2.3; see Ref. 29), and yet HRG disallowed heparin binding by AT in the presence of micromolar amounts of zinc.
Effect of Zinc on Binding of HRG to Immobilized Heparin-Results of the above experiments indicated that supplementary zinc shortened thrombin times of plasmas as well as purified component clotting assays, suggesting that the mechanism for this shortening of thrombin times was zinc-induced HRG binding of heparin. We thus asked if heparin-bound HRG was bound more tightly in the presence of zinc by studying the molarity of NaCl required to elute HRG from an affinity column of immobilized heparin in the presence and absence of zinc. The results of these experiments revealed that 0.1 M NaCl removed a small amount of HRG in 0.1 mM zinc from the column and that 1.0 M NaCl was required to remove most of the HRG from heparin in the presence of zinc (Fig. 7A). In contrast, 0.1 M NaCl was sufficient to remove HRG from heparin in the absence of zinc, as demonstrated by elution in the presence of EDTA (Fig. 7B), which chelates and removes free zinc from solution. These data support the interpretation that zinc significantly increases the ability of HRG to bind heparin and that such binding deprives AT of its heparin cofactor for the inhibition of thrombin.
Comparison of Zinc, Copper, and Magnesium for Heparin Neutralization in the Purified Component Clotting Assay-The specificity of zinc-induced HRG-binding of heparin was studied by using the purified component clotting assay to determine the possible role of other divalent cations reported to bind HRG (4,5,29). The thrombin times of heparinized and non-heparinized assays were determined with and without 5 M zinc chloride, 5 M copper chloride, or 5 M magnesium chloride. The results of these experiments clearly showed that HRG effectively neutralized heparin only in the presence of zinc (Fig.  8), in which case it produced thrombin times that closely approximated those obtained from non-heparinized samples. In addition, control experiments performed by using the purified component clotting assay revealed that the small concentrations of divalent cations used in these experiments were not adequate to precipitate or denature any protein components of the assay (data not shown).

DISCUSSION
Understanding of the function of HRG was broadened by earlier observations that HRG-depleted plasmas were not coagulable in the presence of heparin, while thrombin times of normal plasmas increased only slightly with added heparin (2)(3)(4)(5). We extended these observations in the presence of heparin and found that HRG-supplemented plasmas as compared with control plasmas were not distinguishable by measuring thrombin times, but the effects of increasing HRG concentrations on heparin neutralization became apparent in the pres-

FIG. 7. Elution profiles (280 nm) of HRG from immobilized heparin in the presence and absence of zinc.
A, in the presence of zinc, a small amount of HRG is removed by 0.1 M NaCl (first peak), and 1.0 M NaCl is required to remove the bulk of HRG (second peak). B, in the presence of zinc, a small amount of HRG again is removed by 0.1 M NaCl (first peak), but, in the absence of zinc, 0.1 M NaCl is sufficient to remove the bulk of HRG (second peak). ence of micromolar concentrations of zinc. By using purified component assays, we found that HRG in the presence of zinc bound coagulantly active heparin more effectively than antithrombin. We also found that heparin binding by HRG was increased strikingly in the presence of zinc when studied by heparin-agarose affinity chromatography. Thrombin times were used to test other divalent cations known to bind HRG (4, 10 -12), and zinc was found to be the only divalent cation tested that significantly promoted heparin binding by HRG. Thus, zinc-induced heparin binding is specific, and the cofactor function for HRG binding of heparin in the presence of antithrombin apparently is restricted to divalent zinc.
The amount of zinc required to mediate heparin neutralization in thrombin time assays of citrated plasmas was found to be higher than the amount required for heparin neutralization of heparinized plasmas due to interaction between zinc and the citrate anticoagulant. Specifically, free zinc was not available until citrate binding was complete. This was shown in our experiments performed by utilizing heparinized plasmas and heparinized/citrated plasmas, for the heparinized plasmas required approximately five-fold less zinc than the heparinized/ citrated plasmas to accomplish complete heparin neutralization as measured by restoration of control thrombin times. Indeed, the availability of zinc as measured by heparin neutralization was found to be related directly to the concentration of citrate in the system. Thus, citrate binds and proportionally reduces the amount of free zinc, thereby increasing the amount of supplemental zinc required for heparin neutralization by HRG.
Heparin binding by high molecular weight kininogen (HK) also has been reported to act in a zinc-dependent manner, but the mechanism for this binding is not well understood (30). HRG shares homology with HK, and both proteins belong to the cystatin supergene family (31). Thus, heparin neutralization in our initial plasma experiments could have been due, in part, to HK. However, the heparin neutralization data obtained by the supplementation of plasma with HRG and zinc, and the results of the purified component clotting assays, showed that zinccatalyzed HRG exerts a heparin-neutralizing function independent of HK.
Analysis of the commercial HRG by silver staining and Western blotting techniques under reducing conditions showed major bands corresponding to previously reported native and cleaved forms of 75, 63, 50, and Ͻ 50 kDa (27). The N-terminal region of HRG shares a high degree of sequence homology with the N-terminal region of antithrombin, as well as critical lysine and arginine residues which are essential for heparin binding (32)(33)(34). The metal-binding properties of HRG have been attributed to a repeating amino acid sequence of Gly-His-His-Pro-His (i.e. the histidine-rich region) at the C-terminal end of the native molecule (12,34). Taken together, these findings suggest that native HRG is required for simultaneous metal and heparin binding. A physiological relevance of such binding is suggested by results of recent studies of heparin-binding proteins that have shown HRG to be the only effective competitor with AT for heparin (18). In addition, when we chromatographically isolated the 75-kDa form of HRG and studied its ability with its zinc cofactor to neutralize heparin activity in a purified component clotting assay, we found that it neutralized heparin more effectively than the commercial HRG which contained fragments, suggesting that zinc-promoted HRG neutralization of heparin is attributable to, yet perhaps not necessarily limited to, the native form (i.e. 75 kDa) of HRG. Whether HRG fragments possess a similar capacity remains to be investigated.
The cytoplasm and alpha-granules of human platelets contain zinc at 30 -60-fold higher concentrations than plasma (35); both resting (36) and activated (37) platelets contain HRG. Indeed, serum zinc levels were experimentally shown to be higher than plasma zinc levels. We propose that platelets activated at the reaction site release zinc ions, which serve a cofactor function by binding plasma and platelet HRG molecules that mediate high affinity binding of heparin. This promotes fibrinogenesis only at the reaction site and only in amounts controlled by the microenvironmental concentration of zinc. The procoagulant effect of HRG would be limited by blood flow dilution of microenvironmental zinc and by plasmin cleavage of native HRG (27). Although HRG in purified systems has been shown convincingly to have a lower heparin dissociation constant than AT (4,38), insolubilized unfractionated and low molecular weight heparins bind more AT than HRG from whole human plasma (18). This apparent discrepancy could relate to the availability of zinc as shown by the effects of citrate or EDTA on heparin binding by HRG (29). Thus, our finding suggest that normally thromboresistant vessels can support site-directed fibrinogenesis when activated platelets release zinc ions that increase the heparin binding of HRG versus AT, which focally and momentarily disallows thrombin inhibition. Thrombin times were determined by using the purified component clotting assay with the Coa Screener cuvette containing fibrinogen, thrombin, AT, HRG either with or without heparin in the absence of supplemental divalent cations or identical amounts of zinc, copper, or magnesium.