Mutations in Autolytic Loop-2 and at Asp 554 of Human Prothrombin That Enhance Protein C Activation by Meizothrombin*

Thrombin acts on many protein substrates during the hemostatic process. Its specificity for these substrates is modulated through interactions at regions remote from the active site of the thrombin molecule, designated ex-osites. Exosite interactions can be with the substrate, cofactors such as thrombomodulin, or fragments from prothrombin. The relative activity of (cid:1) -thrombin for fibrinogen is 10 times greater than that for protein C. However, the relative activity of meizothrombin for protein C is 14 times greater than that for fibrinogen. Mod-ulation of thrombin specificity is linked to its Na (cid:2) -bind-ing site and residues in autolytic loop-2 that interact with the Na (cid:2) -binding site. Recombinant prothrombins that yield recombinant meizothrombin (rMT) and rMT des-fragment 1 (rMT(desF1)) enable comparisons of the effects of mutations at the Na (cid:2) -binding residue (Asp 554 ) and deletion of loop-2 (Glu 466 –Thr 469 ) on the relative activity of meizothrombin for several substrates. Hydrolysis of t -butoxycarbonyl-VPR- p -nitroanilide by (cid:1) -thrombin, recombinant (cid:1) -thrombin, or rMT(desF1) was almost identical, but that by rMT was only 40% of that by (cid:1) -thrombin. Thrombin or M and m M adjusted

In the procoagulant reactions of blood coagulation, ␣-thrombin plays a pivotal role in activating coagulation factors (factors V, VIII, XI, and XIII and fibrinogen) and in stimulating a variety of cells such as platelets, leukocytes, and endothelial cells (1)(2)(3)(4)(5)(6)(7). In anticoagulant reactions, ␣-thrombin binds to thrombomodulin, and the thrombin-thrombomodulin complex activates protein C, the proteinase that is key to shutting down the procoagulant processes (8). Meizothrombin and meizothrombin des-fragment 1 are short-lived intermediates in the activation process of prothrombin to ␣-thrombin; both have enzymatic activity (9 -17). Fig. 1 illustrates the cleavage sites in prothrombin and the structures of ␣-thrombin, meizothrombin, and meizothrombin des-fragment 1. Meizothrombin in its normal interaction with thrombomodulin efficiently activates protein C in vitro, but it cleaves fibrinogen inefficiently (18). Moreover, meizothrombin exhibits only 2% of the platelet aggregation activity of thrombin, and its rate of inhibition by antithrombin III is 43% if that of ␣-thrombin (18,19).
In the ␣-thrombin molecule, both autolytic loop-2 (Glu 146 -Lys 149E , chymotrypsin numbering 1 ; residues Glu 466 -Lys 474 of prothrombin) and the Na ϩ -binding region (Asp 221 -Tyr 225 , chymotrypsin numbering; residues Asp 552 -Tyr 557 of prothrombin) are involved in determining thrombin specificity for fibrinogen clotting and protein C (20 -24). When these regions are deleted or substituted in ␣-thrombin to create a non-Na ϩ -binding thrombin form, fibrinogen-clotting activity is remarkably low, but the protein C activator activity decreases only slightly (22)(23)(24). No systematic examination of the role of the Na ϩbinding region and autolytic loop-2 in meizothrombin has been made. If mutations in these regions are constructed in meizothrombin, investigation of the relative activities of the mutants will enable the relationships between Na ϩ binding and the functional differences between ␣-thrombin and meizothrombin to be determined. Determination of the specificity differences between ␣-thrombin and meizothrombin is made difficult by the autolytic cleavages that occur in prothrombin and meizothrombin. We have constructed recombinant prothrombin molecules with substitutions of Ala for Arg at the sites of autolysis to eliminate the confounding cleavages, as reported previously (18), that permit definitive testing of their specificity for two substrates, protein C and fibrinogen.
We then constructed novel meizothrombin mutants with alterations in the Na ϩ -binding site or with a deletion of autolytic loop-2. We report here that these loop-2-and Na ϩ -binding site-modified recombinant meizothrombin derivatives show moderate to large reductions in fibrinogen-clotting activity, but large enhancement of protein C activator activity.

EXPERIMENTAL PROCEDURES
Materials-Human prothrombin was purified following a published method (25). Human protein C was purified using the following modifications of the prothrombin purification method. The Ba 2ϩ -protein precipitate, after (NH 4 ) 2 SO 4 precipitation (25), was applied to a DEAE-Sepharose column and eluted with a linear gradient of 0 -0.5 M NaCl. Fractions that contained factors IX and X and protein C were applied to heparin-Sepharose and eluted with 0.25 M NaCl. The protein C-containing fractions were applied to a Mono Q column and eluted with a linear gradient of 0 -0.5 M NaCl. The purification of human protein C was analyzed by SDS-PAGE at each step. The prothrombin activator ecarin, * This work was supported in part by Health Sciences Research Grants on Advanced Medical Technology from the Ministry of Health, Labor, and Welfare of Japan. 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.
Construction of Prothrombin Mutants-A full-length human prothrombin cDNA was isolated from a human liver cDNA library using two oligonucleotide primers from human prothrombin to which restriction enzyme sites (underlined in the following sequences) were added. The 5Ј-primer sequence was 5Ј-TTTGAATTCACCATGGCGCACGTC-CGAGGC-3Ј, and the 3Ј-primer sequence was 5Ј-CATTGATCAGTTTG-GAGAGCGGCCGCGGTT-3Ј. An amplified cDNA of 1.9 kb was cloned into pUC19, and the resulting pUC/hPT was sequenced using a DSQ-2000L DNA sequencer (Shimadzu Corp., Kyoto, Japan). pUC/hPT was digested with EcoRI and NotI, and the 1.9-kb cDNA fragment was separated by agarose gel electrophoresis and purified using a QIAquick DNA purification kit (QIAGEN GmbH, Hilden, Germany). This fragment was cloned into a mammalian expression vector (pSecTag, Invitrogen) that contains Myc and His tags at the C-terminal end. The protein product of this construct, i.e. human prothrombin produced with pSecTag/hPT, is designated PT-tag. Prothrombin mutants were constructed from pUC/hPT by a PCR-based site-directed mutagenesis method (28). The sequences of all prothrombin mutants were determined by DNA sequencing. In the mutant PT-RA155, Arg 155 is replaced by Ala (the site between fragments 1 and 2) (see Fig. 1); and in the mutant PT-RA271/284, Arg 271 and Arg 284 are replaced by Ala (see Fig.  1). In the mutant PT-RA155/271/284, Arg 155 , Arg 271 , and Arg 284 are replaced by Ala (29). The Na ϩ -binding site mutants PT-DA554 and PT-DL554 have Asp 554 replaced by Ala or Leu, respectively. In PT⌬466 -469, Glu 466 -Thr 469 was deleted from PT-RA155/271/284. All PT constructs contain Myc and His tag sequences at their C-terminal ends, adding 3.4 kDa to these proteins compared with their hypothetical plasma prothrombin-derived homologs. The Myc tag was used to detect the cloned products by anti-Myc immunoblotting.
Expression and Purification of Recombinant Prothrombin Mutants-Expression constructs were transfected into COS-7 cells using FuGENE 6 (Roche Molecular Biochemicals). The transfected cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum for 24 h. Two cultures were grown: one with 10 g/ml vitamin K and one without. The transfected cells were subsequently transferred to serum-free Dulbecco's modified Eagle's medium, again either with or without 10 g/ml vitamin K, and further incubated for 48 h. Expressed PT-tag secreted into the cultured medium was recovered after centrifugation at 2000 ϫ g for 10 min at 4°C to remove the cells.
The supernatant culture medium was applied to an Ni 2ϩ -nitrilotriacetic acid SF column (8-ml volume; QIAGEN GmbH) equilibrated with 50 mM Tris-HCl (pH 8.0) and 0.5 M NaCl to capture the His-tagged protein. The effluent was directly connected to Ä KTA Explorer 10S. The column was washed with 5 column volumes of the equilibrating buffer and eluted by a two-step gradient (0 -40 and 40 -250 mM imidazole in the equilibrating buffer; flow rate of 2 ml/min). One-ml fractions were collected.
The fractions were dialyzed overnight against Tris-buffered saline (TBS; 25 mM Tris-HCl (pH 8.0), 150 mM NaCl, and 5 mM CaCl 2 ), and the protein concentrations of all fractions were determined with a DC protein assay kit (Bio-Rad). Aliquots of the fractions were diluted in 2ϫ SDS buffer (125 mM Tris-HCl (pH 6.8), 30% glycerol, 2% SDS, 0.2% bromphenol blue, and 5% ␤-mercaptoethanol (for disulfide reduction where noted)). Samples were subjected to electrophoresis on 10% SDSpolyacrylamide gels. The proteins were transferred onto a polyvinylidene difluoride membrane (Finetrap NT-32, Nihon Eido, Tokyo, Japan) and incubated with anti-Myc (Invitrogen) and anti-human PT antibodies overnight at 4°C. The proteins were visualized with a VEC-TASTAIN Elite ABC kit (Vector Labs, Inc., Burlingame, CA) and a POD immunostain set (Wako, Osaka, Japan). All prothrombin mutants were expressed and purified by the same method as PT-tag.
Amino Acid Sequence Determination-N-terminal amino acid sequences of the recombinant prothrombins were determined in Ni 2ϩ -nitrilotriacetic acid-purified fractions. Aliquots of the column fractions were diluted in 2ϫ SDS buffer, subjected to electrophoresis on 10% SDS-polyacrylamide gels, transferred to a polyvinylidene difluoride membrane (ProBlott TM , PerkinElmer Life Sciences), and stained with 0.1% Amido Black 10B in 50% methanol and 10% acetic acid. Protein bands to be sequenced were cut out and washed with deionized water (MilliQ, Millipore Corp). The amino acid sequences were determined with an Applied Biosystems Model 477A-120A amino acid sequencer.
Thrombin Assay of Recombinant Prothrombin-All prothrombin mutants were activated by ecarin, from E. carinatus venom (26,27). In a typical assay, the plasma-derived or recombinant prothrombins (50 nM final concentration) were incubated with ecarin (0.5 nM final concentration) in TBS and 0.1% bovine serum albumin at 37°C in a 96-well microtiter plate. After 15 min, 0.1 volume of 2.5 mM Boc-VPR-pNA was added to the activation mixture, and the initial rate of p-nitroaniline formation was monitored at 405 nm with a kinetic plate reader (Wellreader, Seikagaku Corp.). Products of prothrombin activation (after a 30-min incubation) were analyzed by SDS-PAGE and immunoblotting with anti-Myc antibody.
Clotting Assay-Plasma and recombinant prothrombins were adjusted to yield protein concentrations with equal rates of hydrolysis of Boc-VPR-pNA (ϳ0.5 absorbance unit/min) after complete activation with ecarin. Prothrombin samples adjusted to these protein concentrations were incubated with ecarin (0.5 nM final concentration) in TBS and 0.1% bovine serum albumin at 37°C in a coagulometer cup (Amelung KC4A, Baxter). After 30 min, 0.1 volume of 20 mg/ml fibrinogen was added to the activation mixture, and the clotting time was measured. The clotting times were converted to units of thrombin activity using a standard curve constructed from various concentrations of ␣-thrombin from plasma prothrombin.
Protein C Activator Activity-Plasma or recombinant prothrombin (2 nM final concentration) was activated by ecarin (0.2 nM final concentration) in TBS and 0.1% bovine serum albumin at 37°C in a 96-well microtiter plate. After 30 min, equal volumes of human protein C (80 nM final concentration), rTM (either 40 nM final or varying concentrations), and 3:1 (mol/mol) phosphatidylcholine/phosphatidylserine vesicles (100 M final or various concentrations) in TBS containing 0.1% bovine serum albumin were added to the reaction mixture and incubated for 60 min at 37°C to activate protein C. Activated protein C was determined from the initial rate of hydrolysis of Boc-LSTR-MCA (1 mM final concentration), measured by 7-amino-4-methylcoumarin liberation ( ex ϭ 380 nm, em ϭ 465 nm) with a GENios fluorometer (Tecan).
Platelet Aggregation Assay-Platelet aggregation was measured by determining the changes in light transmission of platelet suspensions with an optical aggregometer (Niko Bioscience Hema Tracer 601). Maximal aggregation induced by 100 nM r-␣-thrombin was defined as 100%. Washed platelets were prepared from blood from healthy volunteers with informed consent. All volunteers denied taking any medications that would alter the response of platelets to thrombin. Blood was collected into 0.1 volume of 3.8% sodium citrate by venipuncture and centrifuged at 150 ϫ g for 20 min at room temperature to prepare platelet-rich plasma. Washed platelets were prepared by further centrifugation (300 ϫ g, 3 min) of platelet-rich plasma to separate the platelets, which were then washed twice with 15% acid/citrate/dextrose and 100 nM prostacyclin. Platelet pellets were suspended in HEPES/ Tyrode's buffer (20 mM HEPES (pH 7.4), 138 mM NaCl, 3.3 mM NaH 2 PO 4 , 2.9 mM KCl, 1.0 mM MgCl 2 , 1.0 mM CaCl 2 , and 1 mg/ml glucose). Resuspended platelets were incubated for 2 min at 37°C, after which r-␣-thrombin, rMT, rMT-DL554, or rMT⌬466 -469 (0.1-100 nM final concentrations) was added to initiate aggregation.

RESULTS
To compare the specificity of meizothrombin and meizothrombin des-fragment 1 with that of ␣-thrombin, recombinant prothrombin molecules with a substitution of Ala for Arg at the sites of autolysis were constructed. Fig. 1 shows the cleavage sites and the sites of the mutations; the substitutions and deletions are defined in Table I. PT-tag is an recombinant prothrombin with Myc and His tags at the C terminus, but without other mutations. PT-tag thus serves as the control for recombinant prothrombin and r-␣-thrombin. PT-RA155, with Arg 155 substituted with Ala, is not cleaved to form fragments 1 and 2, but produces r-␣-thrombin that is structurally identical to that derived from PT-tag. PT-RA271/284 and PT-RA155/271/ 284 produce rMT(desF1) and rMT, respectively, and have been reported previously (29). The three mutants PT-DA554, PT-DL554, and PT⌬466 -469 are new constructs designed to produce rMT-DA554, rMT-DL554, and rMT⌬466 -469, respectively. These products permit the evaluation of the contributions of the Na ϩ -binding site Asp residue and autolytic loop-2 to the specificity of meizothrombins for fibrinogen and protein C.
Expression in COS Cells and Purification of Recombinant Prothrombins-The seven recombinant prothrombin constructs were transfected into COS cells. All prothrombin mutants were efficiently expressed ( Fig. 2A). The importance of ␥-carboxylation for prothrombin secretion by COS cells was also demonstrated. ␥-Carboxylation, which modifies a glutamate in the ␥-carboxyglutamic acid domain, depends on vitamin K as a cofactor. Recombinant prothrombin derivatives in COS cells were efficiently secreted into the culture medium only in the presence of vitamin K (Fig. 2B). Thus, prothrombin secretion in this cell system depends on ␥-carboxylation.
The mutants PT-tag and PT-RA271/284 had the molecular masses expected for wild-type prothrombin plus the mass of the tag (Fig. 2, A and B). For recombinant prothrombins with the Arg 155 -to-Ala mutation, the molecular masses appeared higher upon SDS-PAGE by ϳ3-4 kDa (Fig. 2A). The Nterminal amino acid sequences of all expressed recombinant prothrombins were identical to that of wild-type prothrombin (Ala-Asn-Thr-Phe-Leu-). All mutants bound to an Ni 2ϩnitrilotriacetic acid column (Ni 2ϩ -chelating bead), implying that the His tag at the C terminus was correctly constructed and intact. Because both the N and C termini appeared to be correct, the mobility shift of the Arg 155 mutants was not investigated further.
The prothrombin precursor mutants with Myc and His tag sequences at their C termini were purified using metal-chelating column chromatography and detected by anti-Myc antibody. The purity of the PT-tag-containing fractions eluted from the column was demonstrated by SDS-PAGE and by immunoblot analysis (Fig. 3). No autolysis products could be detected in the PT-tag-containing fractions. All mutants were purified by identical metal-chelating column chromatographic procedures (data not shown).  1. Activation products of human recombinant prothrombins. Human plasma prothrombin is converted to ␣-thrombin by the cleavage of Arg 271 and Arg 320 by coagulation factor Xa. ␣-Thrombin cleaves between fragments 1 (F1) and 2 (F2) at Arg 155 and in the ␣-thrombin A chain at Arg 284 to produce fragments 1 and 2 and a small N-terminal fragment of the A chain (residues 272-284). The arrows indicate the sites of proteolytic cleavage or the sites of amino acid mutagenesis. The gray shading shows the activated forms of meizothrombin, thrombin, and the recombinant proteins after activation. A and B are the thrombin A and B chains, which are linked by a disulfide bond.

Activation of Prothrombin Mutants and Functional
Characterization of Activated Forms-As anticipated, PT-tag and PT-RA155 were readily converted to r-␣-thrombin by the prothrombin activator ecarin, whereas PT-RA271/284 and PT-RA155/271/284 were activated to rMT(desF1) and rMT, respectively (Fig. 4). PT⌬466 -469, PT-DA554, and PT-DL554 were similarly converted to the mutant meizothrombin derivatives. All recombinant prothrombins were rapidly activated to their cleaved forms by ecarin, implying that the structure around the relevant scissile bond, Arg 320 , was not grossly perturbed.
We assessed the activity of each recombinant thrombin and meizothrombin derivative by measuring Boc-VPR-pNA hydrolysis rates and by fibrinogen clotting (Table II). PT-RA155 and r-␣-thrombin produced from PT-tag had activities for both substrates experimentally indistinguishable from those of ␣-thrombin derived from native prothrombin. The Boc-VPR-pNA hydrolytic activity of rMT(desF1) was 94% of that of r-␣-thrombin, but rMT activity was decreased to 38%. This is consistent with a published report that the prothrombin fragments, which remain covalently linked in rMT, reduce ␣-thrombin hydrolysis of peptide pNA (30). rMT-DA554, rMT-DL554, and rMT⌬466 -469 had ϳ20% Boc-VPR-pNA hydrolytic activity compared with r-␣-thrombin (Table II). These reductions in activity are similar to those cited in Ref. 31 for different Na ϩ -binding site mutants of ␣-thrombin. Fig. 5 shows the fibrinogen-clotting activities of various activated forms of recombinant prothrombin derivatives. ␣-Thrombin and r-␣thrombin derived from PT-tag had essentially the same clotting activity, but the activity of rMT(desF1) derived from PT-RA271/284 decreased to 16% and that of rMT derived from PT-RA155/271/284 decreased to 12% of the r-␣-thrombin activity. rMT-DA554 and rMT-DL554 had 8 and 6% of the r-␣thrombin activity, respectively. rMT⌬466 -469 had no significant fibrinogen-clotting activity (Ͻ1%). Thus, deletion of autolytic loop-2 abolishes meizothrombin activity for fibrinogen, and mutation of the Na ϩ -binding residue reduces this activity.
Protein C Activator Activity-We examined each of the forms of thrombin and meizothrombin for their ability to activate protein C in the presence of phospholipids and thrombomodulin (Fig. 6A). The significance of ␥-carboxyglutamic acid domaincontaining fragment 1 can be assessed by comparing the activity of rMT(desF1), which lacks fragment 1, with that of the fragment 1-containing rMT derivatives (Fig. 6A, black bars). The activities of rMT and its derivatives were dramatically higher for protein C than those of both rMT(desF1) and r-␣thrombin, in contrast to the lower activities of rMT and its derivatives for fibrinogen. In the absence of covalently attached fragment 1, no enhancement of protein C activation was observed, i.e. with rMT(desF1) or r-␣-thrombin (Fig. 6A). Fig. 6  and C) shows the effects of rTM and phospholipid concentration on protein C activation by r-␣-thrombin, rMT, and rMT derivatives and verifies that the observations are not unique to a limited set of reactant concentrations.
Platelet Aggregation Activity-␣-Thrombin is a potent physiological platelet activator. The effects of r-␣-thrombin and rMT derivatives on platelet aggregation are shown in Fig. 7. Whereas r-␣-thrombin at 3 nM induced 95% platelet aggrega-tion, rMT at 100 nM induced only ϳ60% aggregation. Both rMT-DL554 and rMT⌬466 -469 had poor platelet aggregation-inducing activity (Ͻ5%), even at the highest concentration tested (100 nM). This result shows that both meizothrombin derivatives rMT-DL554 and rMT⌬466 -469 lack significant platelet aggregation-inducing activity. Ecarin, the snake venom-derived prothrombin activator present in these reaction mixtures, did not initiate platelet aggregation (data not shown). DISCUSSION Meizothrombin is known to have enhanced protein C activator activity and diminished fibrinogen-clotting activity compared with ␣-thrombin (11,18,32). In fibrinogen clotting, prothrombin fragment 2 in meizothrombin and meizothrombin(desF1) inhibits interaction with fibrinogen (32). The marked stimulation of meizothrombin protein C activator activity by phospholipids clearly indicates the importance of fragment 1, which includes the Ca 2ϩ -and phospholipid-binding ␥-carboxyglutamic acid domain (33,34), for enhancement of the reaction by phospholipid vesicles (18). Côté et al. (18) observed no enhancement of meizothrombin activity for protein C when soluble thrombomodulin was used rather than the lipidbinding native form of thrombomodulin. One plausible explanation for this difference is the contribution that binding to phospholipid membranes makes to the reaction rates by promoting the formation of thrombin-or meizothrombinthrombomodulin complexes. More complex alterations in the conformation of meizothrombin can also be speculated to occur, e.g. as a result of an intramolecular interaction involving the fragment 1 region and a thrombin exosite. Such conformational changes might be similar to the conformational changes that are caused by thrombomodulin in ␣-thrombin that convert it from a fibrinogen-preferring to a protein C-preferring proteinase. a Relative anticoagulant potency of a mutant, P/F, where P and F are the rates for protein C activator and fibrinogen-clotting activities, respectively. Maximal activities given by r-␣-thrombin are defined as 100%.
b APC, activated protein C; ND, not determined. In ␣-thrombin, deletion of autolytic loop-2 (Glu 146 -Lys 149E , chymotrypsin numbering; Glu 466 -Lys 474 of prothrombin) or the substitution of the Na ϩ -binding site residues (Asp 221 -Tyr 225 , chymotrypsin numbering; Asp 552 -Tyr 557 of prothrombin) decreases its fibrinogen-clotting activity, but affects only slightly its protein C activator activity in the presence of thrombomodulin and phospholipids (20,23,35). In particular, deletion of the entire loop-2 of ␣-thrombin nearly eliminates clotting activity, but protein C activator activity is reduced by only 2-fold (23). Most of the effect must originate from direct perturbation of fibrinogen binding due to the loss of critical interactions between loop-2 and fibrinogen in the autolytic loop-2 deletion mutant.
In this study, Ala or Leu substitution at the Na ϩ -binding site (Asp 554 ) or deletion of autolytic loop-2 residues (Glu 466 -Thr 469 ) ( Fig. 1) in meizothrombin was used to test the importance of the Na ϩ -binding site and autolytic loop-2, known "specificity determinants" for ␣-thrombin, in meizothrombins. The results demonstrate that these regions are also specificity determinants for meizothrombin, but with opposite effects in meizothrombin and ␣-thrombin. Whereas autolytic loop-2 and the Na ϩ -binding residue Asp 554 are crucial for ␣-thrombin action on fibrinogen, mutation of the Na ϩ -binding site in meizothrombin causes only a modest decrease in action on fibrinogen.
The differences in the relative specific activity of the activated meizothrombin mutants for fibrinogen and protein C are more dramatic; in fact, they change from reduced to enhanced relative activity for these two substrates. Specifically, in fibrinogen clotting, rMT is substantially decreased to 12% and rMT-(desF1) to 16% relative to r-␣-thrombin ( Fig. 5 and Table II). rMT-DA554 (8%) and rMT-DL554 (6%) had even lower clotting activity; rMT⌬466 -469 had Ͻ1% of the r-␣-thrombin clotting activity. Because these rMT derivatives (rMT-DA554, rMT-DA554, and rMT⌬466 -469) have no Na ϩ -binding site, their conformation is possibly similar to the non-Na ϩ -binding form of thrombin (35).
The control mutants (r-␣-thrombin from PT-tag and rMT-(desF1)) were indistinguishable from native ␣-thrombin in their ability to catalyze protein C activation in the presence of thrombomodulin and phospholipid vesicles (Fig. 6A). The pro- tein C activator activities of both rMT and its derivatives were increased by 10-fold or more compared with those of ␣-thrombin (Fig. 6A). These four highly active proteins are distinguished by having the fragment 1 region attached, whereas ␣-thrombin, r-␣-thrombin, and rMT(desF1) do not. Both rTM and phospholipids increased the protein C activator activities of the meizothrombin mutants with fragment 1 (Fig. 6, B and  C). Interestingly, even in the absence of thrombomodulin, the fragment 1-containing proteins showed enhanced protein C activator activity compared with ␣-thrombin (Fig. 6A). Table II summarizes the relative chromogenic, fibrinogen-clotting, and protein C activator activities of the activated forms of recom-binant prothrombins. Most significant is the fact that, although rMT-DA554, rMT-DL554, and rMT⌬466 -469 lack an Na ϩbinding function, these activated recombinant meizothrombins have very high protein C activator activities, indicating that the Na ϩ -binding site has an insignificant role (if any) in protein C activation by meizothrombin.
The other procoagulant process recognized as a key function of ␣-thrombin is activation of platelets. In this process, similar to the action of these various forms of thrombin on fibrinogen, the mutations resulted in reduction in relative activity. Fig. 7 shows the effects of r-␣-thrombin, rMT, and meizothrombin derivatives on the aggregation of washed platelets. Both rMT-DL554 and rMT⌬466 -469 had essentially no platelet aggregation activity. These results imply that rMT⌬466 -469 has little or no protease activities for fibrinogen or a physiological substrate such as PAR-1 (protease-activated receptor-1), the thrombin receptor on platelets. Further studies are required to determine whether this is due to decreased binding or decreased ability to catalyze proteolytic cleavage.
It is useful to attempt to relate these observations to the crystal structure of thrombin, although it must be noted that surface residues can be subject to crystal packing variations that prevent detailed interpretations from being made. Fig. 8A shows a model of ␣-thrombin derived from one of the thrombin crystal structures (36). Active-site residues are shown in red. The Na ϩ -binding site (Asp 552 -Asp 554 ) and the autolytic loop-2 region (Glu 466 -Thr 469 ) are yellow and orange, respectively. Other side chains of amino acids in dark green are the interaction sites of fibrinogen, which are located throughout the areas surrounding both the Na ϩ -binding site and the loop-2 region of thrombin (24,(37)(38)(39). rMT-DA554, rMT-DL554, and rMT⌬466 -469 lack a functional Na ϩ -binding site in the thrombin region shown in Fig. 8A. Moreover, rMT⌬466 -469, which is defective in autolytic loop-2, is missing Glu 466 (Glu 146 , chymotrypsin numbering), which ion pairs with Arg 553 (Arg 221A , chymotrypsin numbering) of the Na ϩ -binding site (31,40,41), and had essentially no fibrinogen-clotting activity, suggesting that the side chains of loop-2 might have a significant role in the interaction between meizothrombin and fibrinogen (Fig. 8B). The mutation of Asp 554 (Asp 222 , chymotrypsin numbering), adjacent to the salt-bridging Arg 553 (Arg 221A , chymotrypsin numbering), could similarly be related to the conformational changes that are responsible for the differences between fibrinogen and protein C cleavage by the mutant meizothrombins. In conclusion, it is proposed that interactions within thrombin that involve autolytic loop-2 and the Na ϩ -binding site primarily enhance thrombin action on fibrinogen, but impair thrombin action on protein C.