The Binding of Low Molecular Weight Heparin to Hemostatic Enzymes*

A low molecular weight preparation of porcine hep- arin (specific anticoagulation activity = 125 units/mg) was fractionated to obtain a mucopolysaccharide prod- uct of 6500 daltons (specific anticoagulant activity = 373 units/mg) that is homogeneous with respect to its interaction with antithrombin. This material was treated with fluorescamine in order to introduce a fluorescent tag into the mucopolysaccharide. Initially, we showed that the fluorescamine-heparin conjugate and the unlabeled mucopolysaccharide interacted with antithrombin in a virtually identical fashion. Subsequently, we demonstrated that labeled heparin could be utilized in conjunction with fluorescence polarization spectroscopy to monitor the binding of mucopoly- saccharide to thrombin, factor IXa, factor Xa, and plasmin. The interaction of this complex carbohydrate with thrombin exhibited a stoichiometry of 2:l with a& = =& = 8 X 10" M. The formation of mucopol-ysaccharide~factor IXa complex is characterized by a stoichiometry of 1:l with mt = 2.58 X 10" M. The binding of heparin to factor Xa or plasmin occurred with low avidity. Therefore, the stoichiometries of these processes could not be established. However, our experimental data were compatible with a single-site binding residue with m& = 8.73 X M and af& = -1 X M, respectively.

A low molecular weight preparation of porcine heparin (specific anticoagulation activity = 125 units/mg) was fractionated to obtain a mucopolysaccharide product of 6500 daltons (specific anticoagulant activity = 373 units/mg) that is homogeneous with respect to its interaction with antithrombin.
This material was treated with fluorescamine in order to introduce a fluorescent tag into the mucopolysaccharide. Initially, we showed that the fluorescamine-heparin conjugate and the unlabeled mucopolysaccharide interacted with antithrombin in a virtually identical fashion. Subsequently, we demonstrated that labeled heparin could be utilized in conjunction with fluorescence polarization spectroscopy to monitor the binding of mucopolysaccharide to thrombin, factor IXa, factor Xa, and plasmin. The interaction of this complex carbohydrate with thrombin exhibited a stoichiometry of 2:l with a& = =& = 8 X 10" M. The formation of mucopol-ysaccharide~factor IXa complex is characterized by a stoichiometry of 1:l with m t = 2.58 X 10" M. The binding of heparin to factor Xa or plasmin occurred with low avidity. Therefore, the stoichiometries of these processes could not be established. However, our experimental data were compatible with a single-site binding residue with m& = 8.73 X M and af& = -1 X M, respectively.
Several enzymes of the hemostatic mechanism bind tightly to heparin-Sepharose (1,2). On this basis, it has been tacitly assumed that these proteins would exhibit a significant avidity for this mucopolysaccharide in solution. Numerous attempts have been made to characterize these interactions by indirect means (3, 4). Unfortunately, these studies have yielded confusing and disparate results (3,4).
In this communication, we describe a simple technique for incorporating a fluorescent label into a well defined preparation of heparin that does not alter the functional properties of this complex polysaccharide. In addition, we demonstrate that the labeled mucopolysaccharide can be utilized in conjunction with fluorescence polarization spectroscopy to monitor the binding of heparin to various hemostatic enzymes. Furthermore, we provide direct estimates of the stoichiometries and dissociation constants of these processes.
* This work was supported by Grant HLB-19131 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Recipient of the Established Investigator Award of the American Heart Association. To whom reprint request should be addressed.

MATERIALS AND METHODS
Chemicals-eAminocaproic acid was purchased from Sigma. Fluorescamine (Fluram) was obtained from Roche. All other chemicals utilized were reagent grade or better.
Column Chromatographic Materials-Sephadex G-100 and Sepharose 4B were supplied by Pharmacia.
Heparin-Heparin of porcine origin was obtained from the Wilson Chemical Co. at an early stage in the manufacturing process and prior to treatment of the mucopolysaccharide with oxidizing agents. It was further purified by cetylpyridinium chloride precipitation (5). Low molecular weight species were prepared from this material by filtering 4 g of the mucopolysaccharide at flow rates of 40 ml/h through a column of Sephadex G-100 (5 X 190 cm) equilibrated with 0.15 M NaCl in 0.01 M Tris-HCI, pH 7.5, and pooling fractions with M, = 6OOO to 8000. This product was concentrated by rotary evaporation and extensively dialyzed against 0.15 M NaCl in 0.01 M Tris-HCl, pH 7.5, prior to use.
Proteins-Streptokinase was obtained from Hoechst-Roussel Pharmaceuticals, Inc., Sommerville, N. J. Bovine serum albumin (radioimmunoassay grade) was purchased from Sigma. Purified Russell's Viper venom was generously provided by Dr. Bruce Furie, Boston, Mass. Human factor XIa was a gift from Dr. Allan Kaplan, Human antithrombin was isolated by heparin-Sepharose chromatography and DEAE-cellulose fractionation according to techniques established in our laboratory (6). Human thrombin was prepared in physically homogeneous form by methods previously reported (7).
Human plasminogen was isolated by lysine-Sepharose chromatography (8). Human plasmin was obtained by addition of the above zymogen to streptokinase a t final concentrations of 3 m g / d and 3000 units/ml, respectively. Subsequently, the reaction mixture was incubated for 15 min at 37°C. The solvent utilized was 25% glycerol and 0.05 M lysine in 0.02 M Tris-HC1, pH 8.1. The resultant enzyme preparation exhibited a single band when examined by sodium dodecyl sulfate-gel electrophoresis in the absence of reducing agents.
Human factor IX was isolated as previously described (9). Human factor IXa was prepared by the addition of purified factor XIa to the above zymogen at final concentrations of 11 p g / d and 330 pg/ml, respectively. The resultant mixture was incubated for 1 h at 37°C. The solvent was 0.50 M NaCl and 0.01 M CaCL in 0.01 M Tris-HC1, pH 7.5. The final product exhibited a single band when analyzed by disc gel electrophoresis and sodium dodecyl sulfate-gel electrophoresis in the absence of reducing agents.
Human factor X was isolated according to a previously reported technique (10). Human factor Xa was generated from this zymogen by admixture of purified Russell's Viper venom with this material at final concentrations of 4.5 pg/ml and 178 p g / m l , respectively. The reaction mixture was incubated for 1 h at 37OC. The solvent was 0.20 M NaCl and 0.01 M CaCls in 0.01 M Tris-HC1, pH 7.5. The enzyme produced was homogeneous when examined by disc gel electrophoresis and sodium dodecyl sulfate-gel electrophoresis in the absence of reducing agents.
Measurements of Protein a n d Mucopolysaccharide Concentrations-Protein concentrations were determined by absorbance measurements a t 280 nm. The extinction coefficients of human thrombin, human antithrombin, human factor IX, human factor X, and human plasminogen were assumed to be 16.2 (I]), 6.5 (12), 13.2 (13), 11.6 (14), and 17.0 (15), respectively. Heparin concentrations were estimated colorimetrically by assay of uronic acid according to the method of Bitter and Muir (16) or by measurement of hexosamine as described by Balazs et al. (17). The relationship between these two parameters and the dry weight of heparin fractions was determined experimentally.
Assays ofBiologic Actiuity-Factor IX (18), factor IXa (18), factor X (19), factor Xa (19), factor XIa (la), plasminogen (20), plasmin (20), thrombin (21), and antithrombin (22,23) were assayed by minor modifications of standard techniques. Gel Electrophoresis-The disc gel electrophoretic procedure of Davis (24) as modified by Rosenberg and Waugh (25) was used to ascertain whether protein preparations were homogeneous with respect to charge. The sodium dodecyl sulfate-gel electrophoretic system of Laemmli (26) was employed to determine whether protein preparations were homogeneous with respect to size. In the latter case, the composition of the matrix was established utilizing 0.27% bisacrylamide and 10% acrylamide.
Fluorescence Spectroscopy-Fluorescence measurements were performed with a Perkin-Elmer MPF-44A spectrofluorometer equipped with thermostated sample compartment, polarization accessory, and differential corrected spectra module. The binding of unlabeled heparin to antithrombin was quantitated by measuring the enhancement in fluorescence of tryptophan residues within the protease inhibitor that were excited at 280 nm and emitted at 330 nm. The dissociation constants for the mucopolysaccharide -protein complexes with associated standard errors were calculated by nonlinear least squares fit of the data as previously described (1).
The interactions of fluorescarnine-heparin with hemostatic system enzymes or antithrombin were examined by monitoring the enhancement in polarization (AP) of labeled mucopolysaccharide as a function of protein concentration. These measurements were conducted a t excitation and emission wavelengths of 390 nm and 500 nm to minimize contributions from high levels of protein. Furthermore, a 430nm cutoff filter was employed in front of the emission monochromator to suppress transmission of stray or scattered light.
Polarization values were computed from the following formula: The symbols VH and LH refer to the magnitude of the vertically and The two dissociation constants for this interaction cannot be directly computed by fitting the above data to an appropriate equation. This difficulty is caused by our inability to obtain a simple algebraic expression which relates the magnitude of spectral signals to the concentrations of reactants and their respective dissociation constants. For this reason, we employed an indirect procedure to extract these parameters from experimental observations. T o this end, we derived a series of equations which accurately describe the binding of mucopolysaccharide to protein (see below). Subsequently, these expressions were numerically evaluated for assumed values of the respective dissociation constants and specified concentrations of reactants. The data obtained were used to construct theoretical plots of polarization for given sets of dissociation constants. Finally, these families of curves were compared to the actual observations by appropriate statistical procedures. In this fashion, we were able to choose that pair of dissociation constants which best fit our polarization data.
The theoretical profiles of polarization were computed by employing the following equation: These theoretical profiles were compared to the actual polarization measurements by weighted least squares analyses (27). The plot that exhibited minimal deviation from our binding data was selected. This allowed us to establish the optimal set of ~1 . k~ and which fit our experimental determinations.

Preparation and Characterization of Heparin
Porcine heparin was chromatographed on Sephadex G-100 as described under "Materials and Methods" and mucopolysaccharide species of approximate M , = 6OOO to 8000 were isolated with specific anticoagulant activities that averaged 125 2 10 units/mg. This material was fractionated by a previously reported technique that is based upon the affmity of heparin species for antithrombin (1). The first cycle of this process was initiated by adding protease inhibitor to the mucopolysaccharide pool at a molar ratio of 0.05. Heparin bound to antithrombin, H(B1), was separated from mucopolysaccharide present in solution, H(UI). The H(BI).inhibitor complex was subsequently processed to obtain heparin free of protein. The second cycle of fractionation was initiated by admixing antithrombin with H(U,) at a molar ratio of 0.05. The resultant solution was handled as described above to isolate H(B2) and H(U2). The third, fourth, fifth, sixth, and seventh cycles of fractionation were subsequently conducted in a like manner. In each instance, the protease inhibitor was added to the pool of heparin generated during the preceding stage of fractionation, the mucopolysaccharide-antithrombin interaction product was isolated, and complex sugar was obtained free of extraneous protein. The molar ratio of antithrombin to heparin utilized during these five cycles were 0.05, 0.05, 0.05, 0.15, and 0.15, respectively.
After completing this procedure, the various heparin species were assayed to determine their mucopolysaccharide masses and biologic potencies. Subsequently, a preparative technique was designed to isolate this homogeneous form of "active" heparin. To this end, antithrombin was added to the mucopolysaccharide pool at a molar ratio of 0.08, heparin bound to the inhibitor was harvested, and the complex carbohydrate was obtained free of contaminating protein. The specific anticoagulant activity of this material is 373 f 15 units/mg (average of 6 large scale fractionations). These preparations were filtered at flow rates of 4 ml/h through a column of Sephadex G-100 (0.55 X 180 cm). The chromatographic matrix was previously equilibrated mucopolysaccharide partitioning at each cycle of fractionation.
These estimates were calculated from the relative amounts of with 0.5 M NaCl in 0.01 M Tris-HCI, pH 7.5, and calibrated with heparin standards of known molecular sue. The various products emerged at an elution volume that corresponded to an approximate M, = 6500 (not shown). This material was utilized in all subsequent experiments.

The Incorporation of a Fluorescent Tag into the Heparin Molecule and the Evaluation of the Functional Properties of the Labeled Mucopolysaccharide
In order to introduce a fluorescent label into the heparin molecule, we treated the mucopolysaccharide with fluorescamine. To this end, 1 volume of a solution consisting of fluorescamine dissolved in acetone at a concentration of 0.3 mg/ ml was rapidly admixed with 3 volumes of a second solution composed of mucopolysaccharide dissolved in water at a level of 7.5 X M. After incubation of the resultant mixture for 2 min at 24"C, it was concentrated by rotary evaporation to reduce the content of acetone and then dialyzed against 0.15 M NaCl in 0.01 M Tris-HCI, pH 7.5.
The fluorescent intensity of this heparin conjugate was quantitated by exciting the mucopolysaccharide at 380 nm and measuring the level of emission at 475 nm. Comparisons with labeled compounds, such as 0-methylserine, suggest that approximately 0.4 group of fluorescamine are incorporated per molecule of heparin. These estimates were confmed by radiolabeling the mucopolysaccharide with ['4C]acetic anhydride prior to and immediately after treatment with the fluorescent dye. The difference in I4C content between the two samples indicated that approximately 0.3 group of fluorescamine is incorporated per molecule of heparin. The labeled mucopolysaccharide was subsequently degraded with nitrous acid at pH 1.5 and the resultant fragments were filtered on columns of P-2 polyacrylamide (28,29). Preliminary data obtained by this technique suggest that the a-amino group of serine, within the linkage region, represents a major site for incorporation of tag (not shown).
To demonstrate that our labeling procedure had not altered the functional characteristics of the complex sugar, we compared the interactions of fluorescamine-heparin and unlabeled mucopolysaccharide with antithrombin.' Firstly, we have analyzed the binding of unlabeled heparin to antithrombin with respect to the stoichiometry and avidity of this process. This examination was initiated by admixing varying concentrations of mucopolysaccharide which ranged from 8 X to 5 x M with antithrombin at a constant level of 1 X M. The solvent was 0.15 M NaCl in 0.01 M Tris-HC1, pH 7.5, and 37°C. Subsequently, the various solutions were excited at 280 nm and their enhancements in fluorescence emission at 330 nm were recorded. It was noted that the maximal observed increase in this parameter was attained at a heparin level of 1.3 X M. Given that the concentration of inhibitor utilized in these studies was approximately 100-fold greater than the K& of this interaction (see below), the maximal observed level of fluorescence represented virtually complete saturation of antithrombin with mucopolysaccharide.
In Fig. h  AF,,,,,) versus the molar ratio of mucopolysaccharide to inhibitor employed. As noted, there is a linear increase in AF/AF,,, until a molar ratio of heparin to antithrombin of 0.8 is reached. Thereafter, this parameter grows a t a slower rate until it attains a relatively constant value when the molar ratio of mucopolysaccharide to inhibitor is 1.5 or greater. The stoichiometry of this process was estimated by constructing a line through the initial portion of the ascending limb of the binding isotherm and determining its intersection with the horizontal asymptote that signifies complete occupancy of inhibitor by mucopolysaccharide. This occurs at a molar ratio of heparin to antithrombin of 1.08. Reaction stoichiometries obtained by filtering the radiolabeled mucopolysaccharide and excess inhibitor on columns of Sephadex G-100 are in excellent accord with the above value (see the accompanying paper in this issue, pp. 10081-10090).
The avidity of unlabeled heparin for antithrombin was determined by quantitating the enhancement in the intrinsic fluorescence of inhibitor a t various levels of added mucopolysaccharide. The concentrations of reactants were similar to those employed above except that the level of antithrombin was reduced to 1 x M. This latter value approximates the

K%.& of this interaction (see below) and is, therefore, optimal
for estimating this parameter (30). Fig. 1B  In Fig. 2 A , we have plotted the enhancement in fluorescence polarization of the various samples normalized to the observed maximal increase in this parameter (AP/AP,,,) versus the molar ratio of antithrombin to fluorescamine-heparin utilized. The stoichiometry of this interaction was calculated as outlined above. The value obtained is equivalent to a molar stoichiometry of inhibitor to labeled mucopolysaccharide of 0.96, which is virtually identical with that observed with unlabeled heparin.
The avidity of fluorescamine-heparin for antithrombin was established by measuring the AP of labeled mucopolysaccharide at varying levels of added inhibitor. The concentrations of reactants were similar to those employed above except that the level of fluorescamine-heparin was reduced to 4.62 X M. Fig. 2B represents Unfortunately, this is necessary given the fluorescent intensity of the labeled heparin. Alternately, the incorporation of an intrinsic label into the mucopolysaccharide may minimally distort the binding of heparin to antithrombin.

M.
The Interactions of Heparin with Enzymes of the Hemostatic Mechanism The data provided above demonstrate that fluorescamineheparin and unlabeled mucopolysaccharide interact in a virtually identical manner with antithrombin. Thereafter, flu- to an increase of approximately 80% and represents virtually complete saturation of fluorescamine-heparin with antithrombin, for reasons outlined above. In Fig. 3A, we have plotted AP/AP,,, of the various reaction mixtures uersus the molar ratios of factor IXa to heparin utilized. Appropriate extrapolation of the linear portion of the binding isotherm as previously described suggests that the stoichiometry of this process corresponds to a molar ratio of enzyme to mucopolysaccharide of 1.05.
The avidity of heparin-factor IXa interactions was determined by measuring the AP of labeled mucopolysaccharide at varying levels of added inhibitor. The concentrations of reactants were similar to those utilized above except that the level of fluorescamine-heparin was decreased to 4.29 X 10" M, which approximates Kg&,p (see below). orescamine-heparin was utilized in conjunction with fluorescence polarization spectroscopy to quantitatively evaluate the binding of mucopolysaccharide to factor IXa, thrombin, factor Xa, and plasmin.
Factor IXa-Mucopolysaccharide Interactions-The binding of factor IXa to fluorescamine-heparin was measured under conditions identical with those employed for examining the interaction of antithrombin with labeled mucopolysaccharide. The stoichiometry of this process was evaluated by admixing varying concentrations of factor IXa that ranged ---represents a nonlinear least squares computer fit of our data to a one-binding site model. The multiple r squared value for the curve fit is 0.99.
Binding of Heparin to Thrombin, Factor IXa, Factor Xu, and Plasmin mixtures. The AP,,,., of this process was experimentally defined by observing that AP reached a constant level when thrombin concentrations of 1.0 X 10"' M or greater were employed. This value represents an augmentation in fluorescence polarization of approximately 150%. During these titrations, thrombin contributed less than 5% of the total emission signal at the highest concentration of enzyme used. In Fig. 4 4 , we have plotted AP/APmaX of the various mixtures versus the molar ratio of thrombin to fluorescamine-heparin employed. As noted, a line drawn through the initial linear portion of the binding isotherm intersects the horizontal asymptote a t a molar ratio of thrombin to fluorescamine-heparin of 1.8. This suggests that a single molecule of enzyme is capable of binding two molecules of mucopolysaccharide (see below). It should be emphasized that the extrapolation technique utilized a horizontal asymptote which corresponds to the enhancement in fluorescence polarization generated by complexes formed between single molecules of fluorescamine-heparin and thrombin. These species are expected to predominate as the concentration of enzyme becomes large relative to that of mucopolysaccharide.
Subsequently, we sought to determine the avidity of this complex polysaccharide for thrombin. To this end, we quantitated the A P of fluorescamine-heparin a t varying levels of added enzyme as outlined earlier. However, the concentration of labeled mucopolysaccharide was reduced to 8.57 X 10" M, which is optimal for evaluating the dissociation constants of this interaction. --represents theoretical values calculated from the two-binding site model that best fits our experimental data. These estimates were obtained by numerical evaluation of Equations V and VI with HJb = Gf& = 8 X 10" M . LMW, low molecular weight. molar concentrations of thrombin employed. Our polarization data are most compatible with a two-binding site model. The best statistical fit was achieved by utilizing Kg& = K:& = 8.0 X lo" M (Fig. 4B). However, our analyses also revealed that variations in the magnitude of Kg& and Kg& of -40% had an insignificant effect upon this statistical fit provided that the sum of these two parameters was held constant at twice their initial value (1.6 X 10P M ) .
We have employed the dissociation constants and interaction model provided above to examine the validity of our procedure for estimating reaction stoichiometry. To this end, a theoretical binding curve was constructed at reactant concentrations identical with those utilized for evaluating this parameter. Extrapolation of the linear portion of this isotherm intersected the horizontal asymptote at a molar ratio of heparin to thrombin of 1.8. Thus, the experimentally defined reaction stoichiometry is identical with that predicted from the two-binding site model.
Our analyses of the heparin-thrombin interaction are based upon the assumption that the binding of each molecule of mucopolysaccharide to enzyme will generate an equivalent spectral signal. This appears quite reasonable in view of the fact that an extrinsic probe is utilized to monitor these events and in light of simple physical considerations which govern the enhancement of fluorescence polarization signals. The validity of this assumption is further bulwarked by the close approximation of our mucopolysaccharide-enzyme stoichiometric ratio to 2. However, complex models in which two or more heparins bind to thrombin and contribute in a differential fashion to the resultant fluorescence polarization signal cannot be completely excluded.
Factor Xu-Mucopolysaccharide Interactions-The binding of factor Xa to fluorescamine-heparin was studied under identical conditions and in a manner similar to other hemostatic enzyme-mucopolysaccharide interactions. Given the low avidity of factor Xa for labeled mucopolysaccharide (see below), enzyme levels in excess of 5.0 X M would be required to saturate the fluorescamine-heparin and unambiguously establish the molar stoichiometry of this process. Unfortunately, it proved impossible to conduct studies under these conditions because high levels of factor Xa contributed significantly to our spectral signal. Therefore, we assumed that this process exhibited a molar stoichiometry of 1  polarization studies were also analyzed in terms of the more complex two-binding site model employed for thrombin-mucopolysaccharide interactions. In this case, the two dissociation constants extracted from our data were similar in magnitude to that provided by the single-binding site model. Plasmin-Mucopolysaccharide Interactions-We attemped to investigate the binding of plasmin to fluorescamine-heparin in a fashion similar to that described for the factor Xa-mucopolysaccharide interaction. However, this serine protease is known to exhibit limited solubility in aqueous media unless small amounts of €-amino acids are present (31). Therefore, the binding of protein to mucopolysaccharide was examined in 0.15 M NaCl and 0.005 M e-aminocaproic acid in 0.01 M Tris-HCl, pH 7.5. Utilizing experimental protocols outlined in the previous section we were able to detect modest alterations in the fluorescence polarization of heparin.
For example, varying concentrations of plasmin that ranged from 0 to 2.0 X M were admixed with labeled mucopolysaccharide at a constant level of 5 X 1O"j M. The various solutions were examined as previously described in order to determine the AP of heparin. This parameter exhibited a small increase at very high concentrations of added enzyme ( respectively. The former constant is in excellent agreement with that previously determined (see above), whereas the latter parameter indicates that albumin binds more tightly to heparin than plasmin. These data suggest that the interaction of mucopolysaccharide with plasmin is probably "nonspecific." In summary, we have utilized a highly discriminating but simple isolation technique based upon the affinity of heparin for antithrombin to fractionate a low molecular weight pool of mucopolysaccharide. This approach allowed us to isolate a subpopulation of heparin with a M , = -6500 that is homogeneous with respect to its interactions with protease inhibitor. The availability of this well defined preparation of mucopolysaccharide prompted us to initiate a detailed examination of the binding of this component to the various enzymes of the hemostatic mechanism. Prior to this study, quantitative techniques were not available to directly monitor these interactions. On the one hand, the aromatic amino acid residues of hemostatic enzymes are not significantly perturbed when these proteins bind to heparin. On the other hand, no spectral transitions have been defined within the mucopolysaccharide that might be employed to detect complex formation. Therefore, we decided to covalently attach an extrinsic label to one of the reactants and utilize this moiety to analyze these processes.
To this end, our carefully fractionated heparin preparation was treated with fluorescamine so that a spectral tag could be incorporated into the mucopolysaccharide. Subsequently, we were able to demonstrate that the presence of this extrinsic label did not significantly alter the interaction of this component with antithrombin. Thereafter, fluorescamine-heparin was employed in conjunction with fluorescence polarization spectroscopy to characterize the binding of mucopolysaccharide to thrombin, factor IXa, factor Xa, and plasmin.
The KG& of these processes varied from -1 X M (plasmin) to 2.58 X 10" M (factor IXa) but never attained a value equivalent to K:1ts of 5.72 X lo-' M. Thus our results contradict previous claims that heparin may bind more tightly to hemostatic enzymes than to antithrombin (32, 33). Two of these interactions also exhibited a relatively high KE& that was greater than the similarly designated parameter obtained with albumin. Given the high concentrations of the latter component normally present within blood, it seems unlikely that mucopolysaccharide. serine protease complex formation are required for the anticoagulant action of heparin. However, the binding parameters established during the present investigation will be utilized in the subsequent communication (34) to conduct a detailed evaluation of the kinetic importance of the various heparin-protein complexes.
The approach described in this report should also prove useful in examining other biologic interactions of this type. It is thought that mucopolysaccharides bind tightly to proteins such as fibronectin (35), lipoprotein lipase (36), lipoproteins, collagen, P-thromboglobulin, platelet factor four (37), etc. Unfortunately, it has not been possible to characterize the stoichiometries, specificities, and avidities of these processes. Studies of several of these events utilizing the fluorescence polarization technique outlined above are currently underway in our laboratory.