Isolation and Functional Characterization of the Active Light Chain of Activated Human Blood Coagulation Factor XI*

Human blood coagulation Factor XIa was reduced and alkylated under mild conditions. The mixture con- taining alkylated heavy and light chains was subjected to affinity chromatography on high M, kininogen- Sepharose. Alkylation experiments using [ 14C]iodoa-cetamide showed that a single disulfide bridge between the light and heavy chains was broken to release the light chain. The alkylated light chain (M, = 35,000) did not bind to high M, kininogen-Sepharose while the heavy chain (Mr = 48,000), like Factors XI and XIa, bound with high affinity. The isolated light chain re- tained the specific amidolytic activity of native Factor XIa against the oligopeptide substrate, pyroGlu-Pro- Arg-p-nitroanilide. K,,, and kOatvalues for this substrate were 0.56 mM and 350 s” for both Factor XIa and its light chain, and the amidolytic assay was not affected by CaCl,. However, in clotting assays using Factor XI-deficient plasma in the presence of kaolin, the light chain was only 1% as active as native Factor XIa. Human coagulation Factor IX was purified and la-beled with sodium [‘Hlborohydride on its carbohydrate moieties. When this radiolabeled Factor IX was mixed with Factor XIa, an excellent correlation was observed between the appearance of Factor IXa clotting activity and tritiated activation peptide that was soluble in cold trichloroacetic CA) and 0.5 ml of H20 and counted in a Beckman LS230 liquid scintillation counter for 10 min. To determine the assay back-ground, 24 p1 of the enzyme solution was replaced by 24 pl of TBS containing 10 mg/ml of BSA. In each experiment the assay back-ground was less than 0.6% of the total counts present in the mixture. To determine the maximum release of trichloroacetic acid-soluble tritium from 3H-Factor IX, 24 pl of Factor XIa (136 pg/ml) was employed, and in the presence of 5 mM CaCI2, 52% of the tritium was soluble in trichloroacetic acid. Validation of the activation peptide release assay as an accurate method to determine Factor IX activation was performed by incubating 3H-Factor IX in the presence of Factor XIa. To 3H-Factor IX (2.57 clotting units/ml) in 1.5 ml of TBS/BSA containing 0.13 mM EDTA was added 60 pl Factor XIa (3.5 clotting units/ml) and 200 pl of 50 mM CaC12. The mixture was incubated for 60 min at 37 “C. One hundred seventy-five-pl aliquots were withdrawn as a function of time into 1.4-ml conical centrifuge tubes containing 12 pl of 0.24 M EDTA to stop the reaction. Sixty p1 was analyzed for activation peptide release by trichloroacetic acid precipitation. Ten-to 50-pl aliquots were tested for Factor IXa procoagulant activity in comparison with a standard Factor IXa (a generous gift of Dr. D. L. Aronson, Bureau of Biologics, Bethesda, MD). In a siliconized glass tube, 100 pl of Factor IX-deficient plasma was mixed with 50 p1 of inosithin (1.4 mg/ml in 0.15 M NaCI) and 5 pl of thrombin (4.0 clotting units/ml). The mixture was incubated for 3 min at 37 “C before addition of 50 pl of Factor IXa standard (0.01 to 1.0 clotting units/ml) or sample was added, followed by 75 pl of 50 mM CaC12, and the clotting time was recorded. The Factor IXa procoagulant activity generated upon incubation of 3H-Factor IX with Factor XIa strongly correlated with the release of trichloroacetic acid-soluble radioactivity.

ing site for high M, kininogen. Furthermore the heavy chain region of Factor XIa plays a major role in the calcium-dependent mechanisms that contribute to the activation of Factor IX.
Human blood coagulation Factor XI participates in the early or contact phase of blood coagulation (1). A deficiency of Factor XI can result in excessive bleeding after injury and minor surgery (2). Factor XI circulates in plasma in an inactive zymogen form. The reactions that lead to the activation of Factor XI are initiated upon exposure of blood to negatively charged surfaces. They involve the reciprocal proteolytic activation of Factor XI1 and prekallikrein on the negatively charged surface (3)(4)(5)(6)(7). Activated Factor XI1 then activates Factor XI by limited proteolysis (8,9). In this reaction high M, kininogen serves as a nonenzymatic cofactor (10,11). High M, kininogen is thought to be responsible for binding Factor XI to surfaces adjacent to surface-bound Factor XIIa,' thereby facilitating the action of Factor XIIa on Factor XI (10,13). Stimulated platelets are also able to promote the activation of Factor XI (14)(15)(16) and possess high affinity receptors for Factor XI (17).
Human Factor XI has an apparent M, of 160,000 and consists of two very similar or identical polypeptide chains that are held together by one or more disulfide bonds (8,9,18,19). During the activation of Factor XI by Factor XIIa or trypsin an internal peptide bond in each of the two chains is cleaved giving rise to a pair of disulfide-linked heavy and light chains with M , values of 48,000 and 35,000, respectively (8,9,20). Studies using diisopropylphosphorofluoridate or antithrombin I11 showed that each of these inhibitors bound to the light chain of Factor XIa in a stoichiometry of 2 mol of inhibitor/l mol of enzyme ( M , = 160,000), suggesting that each light chain of Factor XIa bears an active site (9).
Factor XIa is the activator of Factor IX in the intrinsic pathway of blood coagulation (1,21). This reaction occurs in a calcium-dependent two-step mechanism. Initially, an internal peptide bond in Factor IX is cleaved, giving rise to a twochain disulfide-linked inactive intermediate. This intermediate is then converted to Factor IXa by a second cleavage due to Factor XIa, resulting in the release of an activation peptide (21,22). This paper describes studies of the functional roles of the Function of Heavy and Light Chains of Coagulation Factor X I a heavy and light chain regions of Factor XIa. First, we isolated the alkylated light chain of Factor XIa using a method recently developed for this purpose for human plasma kallikrein (23). Then the isolated alkylated light chain was compared to native Factor XIa in terms of its activity on the oligopeptide substrate, pNA, its procoagulant activity, and its activity on 3H-Factor IX.

MATERIALS AND METHODS
All chemicals were the best grade commercially available.
Purification of Proteins-Factor XI and high M, kininogen were purified from human plasma using previously published methods' (24). Factor XI was purified' to greater than 95% homogeneity and contained 250 clotting units/mg of protein when the protein content was determined using the Lowry method (25) with a BSA standard. High M, kininogen was analyzed on SDS-polyacrylamide gels in the presence or absence of reducing agents and migrated as a single polypeptide chain of apparent M, = 110,000. Its specific coagulant activity was 14 clotting units/mg of protein.
Factor IX was purified from commercial Factor IX concentrate. Twelve bottles of Proplex (Hyland Therapeutics, Glendale, CA, Lot 958-C205) were taken up in 10 ml each of 0.05 M phosphate, pH 5.9, containing 40 mM benzamidine, 0.02% NaN,, 10 mM EDTA, and 75 mM NaCl (final 18.5 mmho). These were pooled and dialyzed extensively in the same buffer and loaded on a DEAE-Sephadex A-50 column (2.5 X 15 cm) at 30 ml/h. The resin was extensively washed to elute Factor VII. Factors IX and X and prothrombin were eluted in a single peak with a salt gradient. The terminal buffer was the phosphate buffer described above but containing 0.5 M NaCl (49 mmho). Fractions containing Factors IX and X were pooled and dialyzed against 0.05 M imidazole-HC1, pH 6.0, containing 40 mM benzamidine, 0.02% NaN3, and 0.2 M NaCl(24 mmho). The dialyzed materia1 was made 2.5 mM in CaCI2 and loaded at 30 ml/h on a heparin-agarose column (2.5 X 15 cm) equilibrated in the same buffer containing 2.5 mM CaC1' . The column was extensively washed at 30 ml/h to elute Factor X. Factor IX was then eluted with a salt gradient using this same buffer containing 0.4 M NaCl (39 mmho) as terminal buffer. Six-ml fractions were collected in tubes containing 100 pl of 1 M EDTA. The Factor IX was more than 95% pure as judged by SDS-polyacrylamide gels and contained 166 clotting units/mg. p-Factor XIIa (M, = 28,000) was prepared by incubation of Factor XI1 with trypsin. @-Factor XI1 was then separated from trypsin using DEAE-Sephadex A-50 chromatography as previously described (26).
Polyacrylamide Gel Etectrophoresis-Gel electrophoresis in the presence of SDS was carried out on 7.5% polyacrylamide gels according to Weber et al. (27). The gels were stained for protein with Coomassie blue R-250.
Amino Acid Analysis-Fifteen-pg samples were hydrolyzed in vacuo with 6 M HCI at 150 "C for 24 h and the hydrolysates were analyzed on a Beckman 121 M amino acid analyzer (28).
Preparation of Factor Xla-Factor XIa was prepared from Factor XI using 8-Factor XIIa. To 7 ml of Factor XI (1.54 mg) in 4 mM sodium acetate, 2 mM acetic acid, 0.15 M NaCl, pH 5.3, was added 0.7 ml of 0.5 M Tris, 0.15 M NaCI, pH 7.4, at 37 "C. The reaction was started by the addition of 59 p1 of p-Factor XIIa (38.4 pg). At various times, aliquots were removed and assayed for Factor XIa amidolytic activity. After 18 h Factor XIa amidolytic activity had reached a maximum plateau. Factor XIa was then separated from 8-Factor XIIa using affinity chromatography on CNBr-activated Sepharose 4B (Pharmacia) to which immunopurified anti-Factor XI1 antibodies had been covalently coupled as described elsewhere (29). Factor XIa amidolytic activity was measured using the oligopeptide substrate pyroGlu-Pro-Arg-pNA (S-2366, Kabi Diagnostica). The specific activity of this Factor XIa preparation was 111.4 pmol m i d mg" when S2366 was used at a concentration of 0.37 mM in 0.09 M Tris, 0.09 M NaC1, 1 mg/ml of BSA (Sigma), pH 8.3. Protein concentration of Factor XIa and its light chain (see below) was calculated from data obtained from amino acid analysis following 24-h hydrolysis. Cys and Trp were not determined and were not considered in calculating protein concentration. These account for 7.0% of the amino acid content of Factor XI (8) and, therefore, the protein concentrations and specific enzyme activities may represent overestimates by approximately 7%. Throughout this paper molar concentrations of Isolation of the Light Chain of Factor XIa-7.8 ml of Factor XIa (1.1 mg) in 4 mM sodium acetate, 2 mM acetic acid, 0.15 M NaCl, pH 5.3, was incubated with 1.3 ml of 0.5 M Tris, 50 mM benzamidine, 25 mM EDTA, pH 8.3. Mild reduction of Factor XIa was then performed by incubation at 37 "C with 118.3 p1 of dithiothreitol (Calbiochem, 1 mg/ml) for 45 min under nitrogen in the dark. The reduced Factor XIa was then alkylated with 62.4 pl of iodoacetamide (Calbiochem, 5 mg/ml) for 45 min under the same conditions as used for the reduction. After dialysis against 0.18 M sodium acetate, 5 mM benzamidine, pH 5.3, the reduced and alkylated Factor XIa was applied at a flow rate of 7 ml/h to a column containing 6 ml of CNBr-activated Sepharose 4B to which 18 mg of high M, kininogen had been covalently coupled. The column was washed with five bed volumes of the dialysis buffer, followed by elution of the bound protein with the same buffer containing an additional 0.5 M NaC1. Fractions of 0.8 ml were collected and analyzed for the presence of protein and Factor XIa amidolytic activity. The fractions containing protein were subjected to 7.5% SDS-polyacrylamide gel electrophoresis.
Incorporation of /'4C]Acetamide into the Light Chain of Factor XIa-3.8 ml of Factor XIa (0.87 mg) was reduced as described above. The reduced Factor XIa was alkylated by incubation at 37 "C in the dark under nitrogen. The alkylation was performed using a mixture consisting of 0.77 pmol of ["Cliodoacetamide (Amersham Corp., 54 mCi/mmol) and 0.07 pmol of unlabeled iodoacetamide. After dialysis against 0.165 M sodium acetate, 0.45 M benzamidine, pH 5.3, the light and heavy chains of Factor XIa were separated as described above. The concentration of light chain was determined based on its specific amidolytic activity as described elsewhere (Fig. 2). Samples were counted in a Beckman LS7500 8 counter in Betaphase scintillation fluid, using a known amount of "C as standard.
Amidolytic Assays of Factor XIa and Its Light Chain-The kinetic parameters, K,,, and kc,, for the hydrolysis of pyroGlu-Pro-Arg-pNA by Factor XIa or its light chain were determined as follows. Ten pl of Factor XIa (2.28 pg/ml) or 10 pl of its light chain (0.70 pg/ml) were added to a 1-cm cuvette containing 490 pl of the substrate solution in 0.09 M Tris, 0.09 M NaCl, 1 mg/ml of BSA, pH 8.3. The initial rate of hydrolysis of duplicate samples was measured spectrophotometrically at 405 nm using a Cary 210 spectrophotometer. Km and kcat values were calculated from Lineweaver-Burk plots (30).
Protein concentrations were determined using acid analysis.
Clotting Assays of Factor Xla and Its Light Chain-The procoagulant activity of Factor XIa or its light chain was determined using Factor XI-deficient plasma (George King Biomedical, Inc., Overland Park, Kansas). Fifty p1 of deficient plasma was preincubated at 37 "C for 30 s with 50 pl of Factor XIa or its light chain in different dilutions in 0.01 M Tris, 0.15 M NaCl, 1 mg/ml of BSA, pH 7.4 (TBS/BSA). Fifty pl of cephalin (Sigma) and 50 pl of kaolin (10 mg/ml) or TBS/ BSA were then added and further incubated at 37 "C for 30 s before the mixture was recalcified with 50 p1 of 0.05 M CaC12 and the clotting time measured. The observed clotting time was converted to clotting units by comparison to the clotting activities of serial dilutions of a standard pool of normal human plasma as described before (8). A linear relationship was observed between the clotting time and the log of the concentration of normal plasma, Factor XIa, or its light chain at the concentrations tested. The lines in such plots were parallel for each procoagulant species.
Assay of 3H-Factor IX Actioation-The rate of release of the 3H-Factor IX activation peptide was followed by measurement of the trichloroacetic acid-soluble activation peptide as described before for bovine Factor X (31), bovine Factor IX (32), and human Factor X (33). Factor IX was tritiated and then subjected to preparative polyacrylamide electrophoresis as previously described (33). 3H-Fa~tor IX contained 50,000 cpm/pg and retained full procoagulant activity.
For the studies reported in Figs Diego, CA) and 0.5 ml of H20 and counted in a Beckman LS230 liquid scintillation counter for 10 min. To determine the assay background, 24 p1 of the enzyme solution was replaced by 24 pl of TBS containing 10 mg/ml of BSA. In each experiment the assay background was less than 0.6% of the total counts present in the mixture.
To determine the maximum release of trichloroacetic acid-soluble tritium from 3H-Factor IX, 24 pl of Factor XIa (136 pg/ml) was employed, and in the presence of 5 mM CaCI2, 52% of the tritium was soluble in trichloroacetic acid. Validation of the activation peptide release assay as an accurate method to determine Factor IX activation was performed by incubating 3H-Factor IX in the presence of Factor XIa. To 3H-Factor IX (2.57 clotting units/ml) in 1.5 ml of TBS/BSA containing 0.13 mM EDTA was added 60 pl Factor XIa (3.5 clotting units/ml) and 200 pl of 50 mM CaC12. The mixture was incubated for 60 min at 37 "C. One hundred seventy-five-pl aliquots were withdrawn as a function of time into 1.4-ml conical centrifuge tubes containing 12 pl of 0.24 M EDTA to stop the reaction. Sixty p1 was analyzed for activation peptide release by trichloroacetic acid precipitation. Tento 50-pl aliquots were tested for Factor IXa procoagulant activity in comparison with a standard Factor IXa (a generous gift of Dr. D. L. Aronson, Bureau of Biologics, Bethesda, MD). In a siliconized glass tube, 100 pl of Factor IX-deficient plasma was mixed with 50 p1 of inosithin (1.4 mg/ml in 0.15 M NaCI) and 5 pl of thrombin (4.0 clotting units/ml). The mixture was incubated for 3 min at 37 "C before addition of 50 pl of Factor IXa standard (0.01 to 1.0 clotting units/ml) or sample was added, followed by 75 pl of 50 mM CaC12, and the clotting time was recorded. The Factor IXa procoagulant activity generated upon incubation of 3H-Factor IX with Factor XIa strongly correlated with the release of trichloroacetic acid-soluble radioactivity.

Amino Acid Composition of Factor XZa and Its Light
Chin-The amino acid composition of Factor XIa and its light chain was determined and is shown in Table I. The amino acid composition of Factor XIa was not significantly

Function of Heavy and Light
Chains of Coagulation Factor XIa different from the composition previously reported for human Factor XI (8).
Comparison of the Amidolytic Activities of Factor X I a and Its Light Chain-In order to determine whether the light chain had retained the activity of native Factor XIa, the ability of these two proteins to hydrolyze the small substrate pyroGlu-Pro-Arg-pNA was compared. Fig. 2 shows Lineweaver-Burk plots of the amidolytic activities of Factor XIa and its light chain against this oligopeptide substrate. The K, for pyroGlu-Pro-Arg-pNA obtained from this plot was 0.56 mM and the k,, was 350 s-'. These data (Fig. 2) indicate that the light chain retained the full amidolytic activity of Factor XIa. The fact that the kinetic parameters for dimeric Factor XIa and the single chain light chain are the same suggests that at least for a small substrate the two active sites of dimeric Factor XIa act independently of one another with no cooperative interactions. The material that adhered to the high M, kininogen-Sepharose column and that was eluted by high salt consisted of heavy chain and higher M, forms. This material displayed very low amidolytic activity, probably due to small amounts of partially reduced Factor XIa, which had bound to high M, kininogen-Sepharose and coeluted with the  Amidolytic activity was determined as described in the legend to Factor IX, a natural substrate of Factor XIa, requires calcium ions for a maximal rate of activation by this enzyme. We, therefore, tested the effect of 5 mM CaC1, on the cleavage of the small substrate, pyroGlu-Pro-Arg-pNA, by Factor XIa or light chain. Table I1 shows that calcium ions had no significant effect on the activity of either of the two enzymes on this oligopeptide substrate.
Procoagulant Activities of Factor XIa and Its Light Chain-The ability of Factor XIa and its light chain to correct the clotting defect of Factor XI-deficient plasma was determined (Table 111). In the presence of kaolin, Factor XIa was 100 times as procoagulant as its light chain, whereas in the absence of kaolin it was 340 times as procoagulant. The presence

TABLE I11 Procoagulant activities of Factor XIa and its light c h i n
Clotting assays were performed as described under "Materials and Methods." One clotting unit per ml is defined as the Factor XI clotting activity contained in 1 ml of pooled normal plasma in a kaolin partial thromboplastin time assay (8  of kaolin had no effect on the ability of Factor XIa to correct the clotting time of Factor XI-deficient plasma, whereas the presence of kaolin increased the activity of the light chain approximately 4-fold (Table 111). Cleavage of the Activation Peptide from 3H-Factor I X by Factor XIu or Its Light Chain-Due to the difference in specific clotting activities observed for Factor Xfa or its light chain it was of interest to determine whether this was caused by a differential ability of these two enzymes to activate Factor IX in studies with purified proteins. Therefore, human Factor IX was purified and tritiated on its carbohydrate, and the tritiated activation peptide release assay described by Nemerson and his colleagues (31, 32) was applied to human Factor IX.
Different amount of Factor XIa or light chain were added to an incubation mixture containing 3H-Factor IX and 5 mM CaC12. The release of the trichloroacetic acid-soluble tritium was determined as a function of time. Between 0.028 nM and 0.25 nM of Factor XIa, the initial rate of release of trichloroacetic acid-soluble tritium from 3H-Factor IX displayed a linear increase with the concentration of Factor XIa (Fig. 3). On a molar basis, Factor XIa was about 2000 times more active than its light chain in the presence of 5 mM CaC12 (Fig.  3). To examine the effect of calcium ions on the activity of Factor XIa or its light chain, the release of the 3H-Factor IX activation peptide was measured in the presence of 5 mM CaC12 or 2 mM EDTA or buffer (Fig. 4). The presence of calcium ions stimulated the activity of Factor XIa approximately 600-fold in this assay. However, the activity of the light chain was decreased by about two-thirds in the presence of calcium ions. The presence of EDTA in comparison to buffer alone had no significant effect on either the activity of Factor XIa or its light chain. When the activities of the two enzymes were compared in the presence of 2 mM EDTA or buffer alone, the light chain exhibited approximately 85% of the activity of Factor XIa on a molar basis (Fig. 4).  I X activation by Factor XIa and its light chain 3H-Factor IX activation peptide release assays were performed at 37 "C as described under "Materials and Methods." Plus or minus signs indicate the presence or absence of cephalin, high M, kininogen, or kaolin. When added, kaolin was present at 0.2 mg/ml, high M, kininogen at 0.68 clotting unit/ml, and cephalin at the dilution recommended by the manufacturer for optimal clotting assays. The Factor XIa concentration was 0.15 pg/ml and the concentration of the light chain was 2.0 pg/ml. The initial rate of Factor IX activation was determined under conditions where less than 20% of the activation peptide had been released. Full procoagulant activity of Factor XI in plasma is not expressed unless high M, kininogen, a negatively charged surface, and phospholipids are present together. It is well known that the negatively charged surface and high M, kininogen participate in the reactions leading to the activation of Factor XI, but it is not known whether they also might participate in the activation of Factor IX by Factor XIa. Similarly it is possible that phospholipids also may contribute to the activation of Factor IX by Factor XIa. Therefore, we examined the effect of cephalin, high M , kininogen, and kaolin on the activity of Factor XIa or its light chain using the 3H-Factor IX activation peptide release assay. As seen in Table  IV, the presence of cephalin or high M , kininogen had no enhancing effect in this assay for either the activity of Factor XIa or its light chain. Kaolin had a slight inhibitory effect on the activity of Factor XIa, either alone or in combination with high M , kininogen (Table IV). Higher concentrations of kaolin in the reaction mixture, such as 2 mg/ml, either alone or in combination with high M , kininogen inhibited the activity of Factor XIa by 96%.

DISCUSSION
This study was undertaken to examine relationships between the functional properties and physical domains of human Factor XIa. Limited proteolysis of Factor XI with p-Factor XIIa results in its activation to form Factor XIa (8,9). Human Factor XIa is composed of four disulfide-linked polypeptide chains, including two heavy chains (Mr = 48,000) and two light chains ( M , = 35,000) (8,9,20). Conditions are described here under which mild reduction and alkylation of Factor XIa may be performed and a fully active light chain enzyme may be recovered. Fig. 1 shows that under the conditions of mild reduction not all of the Factor XIa molecules were completely reduced. Partially reduced forms of Factor XIa were present at M, = 130,000, 100,000, and 80,000. The band at M, = 80,000 probably represents reduced Factor XI and/or two-chain Factor XIa.
The band at M, = 100,000 might represent a dimer of heavy chains, suggesting that interchain disulfide bond(s) of Factor XIa are located between the heavy chain domains. The 130,000 M, band may represent  (9). Hence, it was inferred that both light chains contain an active site. Moreover, an amino acid sequence typical of an active site serine region is located in the light chain (34).
Since Factor XIa and its light chain exhibit identical activities against a small substrate, we next examined the question of their relative activities in the more physiological setting of a clotting assay. The fact that Factor XIa is 100-fold more active than its light chain in a clotting assay (Table 111) indicates that the heavy chain of Factor XIa has a very important role in coagulation reactions. Kaolin had no effect on the procoagulant activity of Factor XIa and it had a 4-fold stimulatory effect on that of the light chain.
One obvious explanation for the role of the heavy chain in expressing the procoagulant activity of Factor XIa in plasma would relate to the interaction of Factor XIa with its natural substrate, Factor IX. In order to explore this possibility, we studied the activation of purified Factor IX by Factor XIa or its light chain using the tritiated Factor IX activation peptide release assay (31-33). In this assay Factor XIa proved to be approximately 2000 times as active as the light chain in the presence of CaC12. The two forms of the enzyme, however, exhibited the same activity against 3H-Factor IX in the absence of CaC12. Interestingly, calcium ions inhibited the activity of the light chain approximately 65%. Factor XIa in the presence of calcium ions was 600 times more active than the light chain or Factor XIa in the absence of calcium ions. These unexpected findings suggest that the heavy chain of Factor XIa is intimately involved in calcium-dependent mechanisms that accelerate the activation of Factor IX.
The role of calcium ions in the activation of Factor IX by Factor XIa has usually been assumed to be confined to interaction of calcium ions with Factor IX (35-37) since Factor IX binds calcium ions and no direct evidence has been published bearing on the interaction of calcium ions with Factor XI or Factor XIa. Human Factor IX binds 16 calcium ions and positive cooperativity was demonstrated for at least the first four sites (36, 37). A slight conformational change may occur upon calcium binding to Factor IX as evidenced by perturbations of tryptophanyl chromophores (37), although no large changes occur in the circular dichroism spectrum (36). Our data may suggest that the heavy chain of Factor XIa has a binding site for calcium ions which is required for a maximal rate of activation of Factor 1X. Another interpretation is that calcium ions are required for Factor IX to be in a conformation that is necessary or optimal for interaction with the heavy chain of Factor XIa. Interaction of Factor IX in the presence of calcium ions with the heavy chain of Factor XIa could accelerate the reaction by enhancing enzyme-substrate interactions. Still there is no direct evidence that the heavy chain of Factor XIa interacts directly with Factor IX in the presence of calcium.
It is possible that differences in the activities exhibited by Factor XIa and its light chain are not due to functional properties provided by the heavy chain in the former, but rather due to a change in conformation of the light chain that occurs upon reduction and alkylation of intrachain disulfide bridges or upon alkylation of residues other than cysteine. These two possibilities are unlikely, since the experiment using ['4C]iodoacetamide showed that only one molecule of acetamide is incorporated per molecule of Factor XIa light chain. This also shows that only one disulfide bridge attaches the light chain to the heavy chain and that the heavy chains are linked to each other by disulfide bonds.
No significance has yet been associated with the fact that Factor XIa is a covalent dimer containing two active sites. Disulfide bonds between the heavy chains of Factor XIa are essential for the dimeric nature of Factor XIa and the possibility of noncovalent interactions between the light chains cannot be excluded. Perhaps the dimeric nature of Factor XIa is important for efficiently cleaving two different peptide bonds in Factor IX, although with small substrates each active site of the dimeric Factor XIa is equally as active as the active site of the monomeric light chain.
The experiments shown in Table IV  Refs. 1 and 39). They each bind reversibly to high M, kininogen, a cofactor in their surface-dependent activation by Factor XIIa. Moreover, Factor XIa as well as kallikrein are potent activators of surface-bound Factor XI1 and weak activators of plasminogen. Recently we reported studies of human plasma kallikrein (23) similar to those in this paper. AS with Factor XIa, the light chain of kallikrein contains the active site and the heavy chain possesses the high affinity binding site for high M , kininogen. The procoagulant activities of the alkylated light chains are greatly decreased although amidolytic activities are fully retained. Unlike the light chain of Factor XIa, however, the light chain of kallikrein retained full proteolytic activity against the macromolecular substrate, Factor XII, in solution. The large difference between the procoagulant activity of kallikrein and its light chain probably reflects the inability of the light chain of kallikrein to activate surface-bound Factor XI1 as effectively as does native kallikrein. 3 The large difference between the procoagulant activity of Factor XIa and its light chain is due to the calcium-dependent cleavage and activation of Factor IX. Thus, the heavy chain regions of Factor XIa and kallikrein not only contribute to the Factor XIIa-dependent activation of these molecules by binding to the surface cofactor, high M, kininogen, but also contribute substantially to the action of these enzymes on their macromolecular substrates of the coagulation pathways.
Factor XIa activates Factor IX in the absence of calcium ions although at a much reduced rate as shown here and previously (36, 38). Kallikrein also activates Factor IX but calcium ions do not accelerate this reaction and kallikrein is 20,000 times less active than Factor XIa (22). Measurements of the release of tritiated activation peptide from 3H-Factor IX by kallikrein and its isolated light chain showed that the light chain of kallikrein was as active as kallikrein in the presence or absence of calcium: Moreover, the light chain of Factor XIa was approximately 10 times as active as the light chain of kallikrein.4 Thus, the heavy chain of kallikrein does not contribute to the ability of that enzyme to active Factor IX and there is no calcium effect on the reaction. These observations reinforce the notion that the heavy chain region of Factor XIa somehow contributes critically to calciumdependent mechanisms that enhance the activation of Factor IX. Since neither phospholipid nor kaolin enhance the activation of Factor IX by Factor XIa, calcium ions must directly enhance enzyme-substrate interactions.