Expression, Isolation, and Characterization of an Active Site (Serine 528-Alanine) Mutant of Recombinant Bovine Prothrombin*

An active site mutant bovine prothrombin cDNA (Serfiz8+Ala) has been constructed, subcloned, and expressed in Chinese hamster ovary cells. The recombi- nant mutant prothrombin, expressed at the level of 1.5-2.0 wg/ml of cell medium, was fully carboxylated (9.9 f 0.4 mol of y-carboxyglutamic acid/mol of pro-thrombin). The mutant prothrombin could be activated to thrombin by Taipan snake venom and activated to meizothrombin by ecarin in a manner comparable to native bovine prothrombin or recombinant wild-type bovine prothrombin. The mutant meizothrombin thus formed was stable and did not autolyze. The initial rate of cleavage of mutant prothrombin catalyzed by the full prothrombinase was only 28% of the rate of cleav- age of native prothrombin, while recombinant wild-type prothrombin was cleaved at the same rate as the native molecule. The mutant thrombin, obtained from the mutant prothrombin in situ by prothrombinase or Taipan snake venom activation, showed no enzymatic activity toward either fibrinogen or a synthetic chro- mogenic substrate, D-phenylalanyl-L-pipecolyl-bar-ginine-p-nitroanilide a X lo-' than = 1.7 X lo-' M). These results indicate that the mutant recombinant prothrombin described a useful tool for the study of meizothrombin or thrombin without the complications arising from the proteolytic activities of these molecules. Study of the activation of this mutant revealed a func-tional link between the site of initial cleavage by the prothrombinase and the conformation at the nascent active site of prothrombin.


Expression, Isolation, and Characterization of an Active Site (Serine 528-Alanine) Mutant of Recombinant Bovine Prothrombin*
Gang Pei, Keith Baker$, S. Michelle Emfinger$, Dana M. Fowlkes$, and Barry R. Lentz$ From the Department of Biochemistry and Biophysics and $Department of Pathology, Uniuersity of North Carolina, Chapel Hill, North Carolina 27599-7260 An active site mutant bovine prothrombin cDNA (Serfiz8+Ala) has been constructed, subcloned, and expressed in Chinese hamster ovary cells. The recombinant mutant prothrombin, expressed at the level of 1.5-2.0 wg/ml of cell medium, was fully carboxylated (9.9 f 0.4 mol of y-carboxyglutamic acid/mol of prothrombin). The mutant prothrombin could be activated to thrombin by Taipan snake venom and activated to meizothrombin by ecarin in a manner comparable to native bovine prothrombin or recombinant wild-type bovine prothrombin. The mutant meizothrombin thus formed was stable and did not autolyze. The initial rate of cleavage of mutant prothrombin catalyzed by the full prothrombinase was only 28% of the rate of cleavage of native prothrombin, while recombinant wildtype prothrombin was cleaved at the same rate as the native molecule. The mutant thrombin, obtained from the mutant prothrombin in situ by prothrombinase or Taipan snake venom activation, showed no enzymatic activity toward either fibrinogen or a synthetic chromogenic substrate, D-phenylalanyl-L-pipecolyl-barginine-p-nitroanilide dihydrochloride (52238). The mutant thrombin also bound dansylarginine-N-(3ethyl-1,5-~entanediyl)amide, a specific fluorescent inhibitor of the thrombin active site, with a weaker binding affinity (kd = 5.4 X lo-' M) than did native thrombin (kd = 1.7 X lo-' M). These results indicate that the mutant recombinant prothrombin described here is a useful tool for the study of meizothrombin or thrombin without the complications arising from the proteolytic activities of these molecules. Study of the activation of this mutant has already revealed a functional link between the site of initial cleavage by the prothrombinase and the conformation at the nascent active site of prothrombin.
Prothrombin is the precursor of thrombin, a serine protease which plays a central role in the process of blood coagulation (Mann et al., 1988). The activation of bovine prothrombin to thrombin requires cleavage of two peptide bonds, Arg3"3-Ile:'24 and Arg274-Thr275. Cleavage at Arg32"-Ile324 results in exposure of the active site and formation of meizothrombin, and, if al., 1981;Rosing and Tans, 1988). Like all other serine pro-* This work was supported by United States Public Health Service Grants HL26309 (Specialized Center for Research in Thrombosis and Hemostasis, to B . R. L.) and HL36197 and HL31012 (to D. M. F.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Arg274-Thr277" has already been cleaved, thrombin (Mann et $ T o whom reprint requests should be addressed. teases, the thrombin active site contains the catalytic triad, composed of His366, Asp4", and Ser5". The activity of thrombin active site toward fibrinogen, factor V, factor VIII, factor XIII, and platelets makes it very difficult to measure the thrombin binding to those substrates without activation. The cleavage of Arg'"-Ser'57 in bovine prothrombin and of both Arg155-Ser156 and ArP6-Thrza7 in human prothrombin by either thrombin or the activation intermediate meizothrombin also creates an obstacle to the crystallization of prothrombin or to biophysical studies of meizothrombin. Meizothrombin undergoes autolysis even in the presence of potent active site inhibitors (Armstrong et al., 1990;Doyle and Mann, 1990;Pei and Lentz, 1991). Therefore, it is desirable to have an active site mutant of prothrombin.
A wild-type human prothrombin cDNA has been expressed in Chinese hamster ovary cells (CHO),' and the expressed recombinant prothrombin has been shown to be equivalent to plasma-derived prothrombin in specific coagulant activity and in y-carboxyglutamic acid content (Jorgensen et al., 1987). A series of point mutations in the amino-terminal region of prothrombin were successively made to identify the amino acid residues in the y-carboxylation recognition site (Huber et al., 1990). In the study reported here, a wild-type bovine prothrombin cDNA and an active site mutant (Ser5"+Ala) bovine prothrombin cDNA have been constructed and expressed in CHO cells. The mutant recombinant prothrombin obtained was fully y-carboxylated, and was activated at similar rates by ecarin and Taipan snake venom and at a slower rate by the full prothrombinase as compared to native prothrombin or wild-type recombinant prothrombin. The mutant thrombin had no activity toward fibrinogen or toward a synthetic chromogenic substrate (S2238) and displayed a moderately decreased binding affinity toward dansy1arginine-N-(3-ethyl-1,5-pentanediyl) amide (DAPA), a specific thrombin active site inhibitor.
["C]Formaldehyde was purchased from Du Pont-New England Nuclear. Insulin-transferrin-sodium selenite media supplement (ITS) and NuSerum'" were from Collaborative Research (Bedford, MA). Cell growth medium Dulbecco's modified Eagle's medium/F-12 and calf serum were obtained from Linberger Cancer Research Center, University of North Carolina, Chapel Hill.
Isolation of bovine prothrombin and preparation of small unilamellar vesicles (SUV) of PS and PC (PS/PC, 25/75) were as described (Tendian and Lentz, 1990). Factor X and factor V were isolated from human plasma by procedures routinely used in our laboratory (Cutsforth et al., 1989). Factor X was activated with a purified fraction of Russel's viper venom factor X-activating protein which had been covalently linked to agarose beads (Jesty and Nemerson, 1976). Factor V was activated by 1.25 ng/ml factor V-activating protein isolated from Russell's viper venom (Kisiel and Canfield, 1981) in situ. Thrombin was generated and purified according to the procedure of Church and Whinna (1986) and was coupled to agarose beads (Affi-Gel 10, Bio-Rad) following the instructions provided by Bio-Rad.
Rabbit antisera were raised against purified bovine thrombin using incomplete Freund's adjuvant. Protein concentrations were determined by absorbance measurements using extinction coefficients a t 280 nm of 1.44 ml mg" cm" for prothrombin and 1.05 ml mg" cm" for prothrombin fragment 1 (Mann, 1976).
Construction of Expression Plasmids, Wild-type and Mutant Bovine Prothrombin cDNA Clones-The expression vector used for these experiments was a derivative of a shuttle vector, pDx, from Dr. K.
Berkner, ZymoGenetics (Seattle, WA). pDx was modified by the insertion of an assembled pair of oligonucleotides (aattcactcgagagcggccgc and aattgcggccgctctcgagtg) into the unique EcoRI site. This placed a unique NotI site immediately downstream of the unique EcoRI site. The resulting plasmid was called p272 (details not shown). New subcloning vectors were also constructed. p287 was derived from pBS(-) (Stratagene, San Diego) by cloning a small oligonucleotide fragment (agcttccatggtcgcgactcgagctgca and gctcgactcgcgaccatgga) encoding a NcoI restriction site between the HindIII and PstI sites. p431 was constructed by inserting an oligonucleotide fragment (agcttgcggccgcctgcaggatatcccatggtggccag amd aattctggccaccatgggatatcctgcaggcggccgca) encoding NotI, PstI, and NcoI sites between the HindIII and EcoRI sites of p287. Two plasmids, pBII-3 and pBII-111, containing amino-and carboxyl-terminal portions of bovine prothrombin were obtained from Ross MacGillivray, University of British Columbia (MacGillivray et al., 1980). The prothrombin-coding sequence containing the initiating methionine codon is located within a unique NcoI restriction enzyme recognition site in pBII-111. To obtain a full-length bovine prothrombin cDNA, we subcloned the 450-base pair PstI-BamHI fragment from pBII-3 and the 1500-base pair NcoI-BamHI fragment into p287 (Fig. 1A). The resulting clone was called p429. The NcoI-PstI fragment of p429, containing the complete prothrombin cDNA was subcloned into p431 to form p433. Hence, p433 contains a unique NotI site at the 3' end of the prothrombin cDNA (Fig. 1A). The prothrombin expression vector, p437 ( Fig. lC), was constructed by subcloning the EcoRI-Not1 fragment of p433 into the shuttle vector p272.
The 500-base pair BamHI-Hind111 fragment from p433 encoding the carboxyl-terminal region of prothrombin was subcloned into pUC-8, resulting in a clone called p479 (Fig. 1B). To mutate the active site serine (Ser"") to alanine, we replace the BamHI-Ban11 fragment of p479 with a fragment assembled from four synthetic oligonucleotides to yield clone p481. The sequences of these oligonucleotides are gatccgcatcaccgacaacatgttctgtgccggttacaagcctggtgaaggc, aaacgaggggac gcttgtgagggcgacgcagggggacccttcgtcatgaagagcc, cttcatgacgaagggtccccc tgcgtcgccctcacaagcgtcccctc, gtttgccttcaccaggcttgtaaccggcacagaacatgtt gtcggtgatgcg. The BamHI-Hind111 fragment of p481 was subcloned into p433, resulting in p485 (see Fig. 1B). The region of Ser5"+Ala was sequenced by the dideoxynucleotide method (Sanger et al., 1977) to confirm the presence of the SerSz8+Ala mutation and the lack of other undesirable mutations. The EcoRI-Not1 fragment of p485 was subcloned into p272. The result is a shuttle expression vector, p487 ( Fig. IC) that has the potential to express Ser5"+Ala prothrombin.
Expression of Mutant and Wild-type Prothrombin in CHO Cells-Calcium phosphate cotransformation of CHO was carried out using a 1:lO ratio of p437 or p487 to pRSVneo according to a standard protocol (Graham and van der Eb, 1973). Cells were grown in Dulbecco's modified Eagle's medium/F-12 with 10% calf serum and 500 pg/ml G418, a neomycin analog (GIBCO Bethesda Research Laboratory). Individual clones of CHO cells resistant to G418 were tested for production of prothrombin. Nearly confluent cells were washed twice in Dulbecco's modified Eagle's medium/F-lP/ITS and 10 pg/ ml vitamin K. Two ml of media was left on each plate for 24 h. The media from the CHO cells transformed with wild-type prothrombin expression vector were tested for their ability to hydrolyze Spectrozyme TH as recommended by the supplier (American Diagnostica, NY) upon activation with ecarin. Six colonies produced significant levels of prothrombin. One of these has been used for the studies presented below. The media from CHO clones producing the mutant form of prothrombin were assayed for production of prothrombin using Western blot analysis with the rabbit antisera. One clone that produced detectable levels of mutant prothrombin has been used for the studies. Stocks of cells were maintained in Dulbecco's modified Eagle's medium/F-12 with 5.0% NuSerum, 5.0% calf serum, and 500 Kg/ml G418.
To scale up cell production, one confluent 100-mm plate was transferred to one roller bottle (Costar, 900 cm') in a total volume of 250 ml. In about 4 days, the cells were confluent. Sterile microcarrier beads (0.8 g, Pharmacia LKB Biotechnology, Inc., Cytodex 2), diluted in 50 ml of growth medium, were added to each bottle. After 3 days, when the beads were covered with cells, they were washed once briefly with serum-free medium and once again overnight. All medium was removed, and each bottle was fed with 250 ml of serum-free medium. About 50 ml of conditioned medium was harvested daily, with 50 ml of fresh medium added, for 3 weeks. A final concentration of 10 mM benzamidine and 0.15 mM phenylmethylsulfonyl fluoride was added into the medium collected.
Isolation of Recombinant Prothrombins-Briefly, 3.7 ml of 0.4 M sodium citrate and 0.1 ml of 1.0 M benzamidine were added to 100 ml of the cell-free medium containing recombinant prothrombin. Prothrombin was precipitated with l/lOth volume of 1 M barium chloride, and the precipit.ate was washed three times with 0.1 M barium chloride and 1 mM benzamidine. The precipitate was resuspended in 40% saturated ammonium sulfate and 0.2 M EDTA and subjected to centrifugation. The supernatant was then dialyzed against 20 mM Tris, 1 mM benzamidine, pH 7.4, overnight and loaded onto a FPLC Mono Q ion-exchange column (HR 5/5, Pharmacia LKB Biotechnology, Inc.). The column was washed with 20 mM Tris, pH 7.4, buffer, and the protein was eluted with a linear gradient from 0 to 0.7 M NaCl in the same buffer a t 1.5 ml/min. Protein absorbance was monitored a t 280 nm. The identity of proteins in different fractions was determined by sodium dodecyl-sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), using prestained SDS-PAGE standards (Bio-Rad) and purified native prothrombin as the standard. The fractions containing pure prothrombin were combined and stored a t -70 'C for the following studies.
y-Carboxyglutamic Acid Content of Recombinant Prothrombins -Prothrombin fragment 1 was generated by cleavage of prothrombin with thrombin coupled to agarose beads. The resulting fragment 1 and prethrombin 1 were separated on a FPLC Mono Q column with a linear gradient of NaCl (0-0.7 M) in 20 mM Tris, pH 7.4, buffer. Taking native prothrombin fragment 1 as the standard, the relative y-carboxyglutamic acid (Gla) content of fragment 1 of recombinant prothrombins was determined according to the Gla-specific formaldehyde-morpholine Mannich base fragmentation method (Wright et al., 1986). This method converts the 10 Gla residues of prothrombin fragment 1 into 10 y-methyleneglutamyl residues and can quantitatively label the Gla residues when [14-C]formaldehyde is used. Protein concentration of the modified fragment F1 was determined by absorbance a t 280 nm and 14C radioactivity of the modified fragment F1 was counted on a LKB 1211 Rackbeta liquid scintillation counter.
Prothrombin Activation by Prothrombinase and Enzymatic Activity of the Activation Products-Mutant recombinant prothrombin, wildtype recombinant prothrombin, and native prothrombin (all at 1.2 p M ) were activated with 0.5 nM factor x,, 5 nM factor v,, 10 p M PSI PC (25/75) SUV and 5 mM CaC1' in 20 mM Tris, 150 mM NaCl, pH 7.4, a t 22 "C. At intervals of 0,5, and 10 min, aliquots were withdrawn and subjected to SDS-PAGE (10% polyacrylamide) under nonreducing and reducing (3% 2-mercaptoethanol) conditions. Gels were stained with Coomassie Blue R-250 (Sigma), and the protein bands were quantitated by densitometry (Molecular Dynamics, Sunnyvale, CA) to estimate the initial rate of prothrombin activation. Activation was also carried out at 37 "C and thereby taken to completion in 45 min. The product was thrombin in every case as shown by SDS-PAGE. An aliquot of each incubation was taken to measure the activity toward synthetic chromogenic substrate 52238 and fibrinogen (Sigma). The measurements were performed simultaneously on a 96-well plate with a Microplate Reader ( The expression vector for the wild type (p437) was made by subcloning the EcoRI-Not1 fragment of p433 ( A ) into p272. The mutant prothrombin (Ser5*'+Ala) expression vector (p487) was made by subcloning the EcoRI-Not1 fragment of p485 (B) into p272 (see "Experimental Procedures" for details). Resulting vectors are shown in C.
chromogenic substrate assay, 5-pl aliquots (equal to 0.45 pg of prothrombin), 60 p1 of 1 mM S2238, and 120 p1 of 20 mM Tris, 150 mM NaCI, pH 7.4 buffer were used. For the fibrinogen clotting assay, 2-p1 aliquots (equivalent to 0.18 pg of prothrombin), 140 p1 of 5 mg/ml fibrinogen, and 50 pl of 20 mM Tris, 150 mM NaC1, pH 7.4 buffer were used, and turbidity was recorded at the same wavelength. Prothrombin Actiuation by Ecarin and Taipan Venom-Recombinant prothrombins were also activated with ecarin from E. carinatus venom, which cleaves the Arg32'-Ile3'4 bond of bovine prothrombin (Rosing and Tans, 1988). Ecarin (12.5 Fg/ml final concentration) was added to a final concentration of 90 pg/ml of native or recombinant prothrombin in 20 mM Tris, 150 mM NaC1, 6 mM CaC12, pH 7.4, buffer a t 37 "C to start the activation. At intervals of 0, 30, 90 min, 20-pl aliquots were withdrawn and subjected to SDS-PAGE under both nonreducing and reducing conditions. Gels were stained with Coomassie Blue R-250.
The DAPA fluorescence assay (Mann et al., 1981;Nesheim and Mann, 1983;Krishnaswamy et al., 1986) was also applied to compare the activation of mutant and native prothrombin by ecarin and Taipan venom. Taipan snake venom from 0. scutellatus is known to activate prothrombin to thrombin in the absence of factor V, a t a rate that is enhanced 20-30 times by phospholipid vesicles (Owen and Jackson, 1973). Using an SLM 48000@ fluorometer with excitation wavelength a t 280 nm and emission wavelengths a t 515-550 nm with a 515-nm cut-off filter, 125 nM mutant or native prothrombin, 5 mM CaC12, and 500 nM DAPA in 20 mM Tris, 150 mM NaCl, pH 7.4, buffer were incubated at 37 "C for 3 min before adding 100 pg/ ml ecarin or 10 pg/ml Taipan venom (final concentrations) to start the reaction. The Taipan venom activation mixture also contained 10 p~ PS/PC (25/75) SUV. The fluorescence intensity of DAPA was greatly enhanced when bound to the active site of newly formed meizothrombin or thrombin, and this allowing us to record continuously the time course of prothrombin activation. The fluorescence emission spectra of both reaction mixtures were also determined from 450 to 620 nm (4-nm bandpass) with an excitation wavelength of 280 nm and a 450-nm cut-off filter in the emission beam.

overnight. Prothrombin was completely activated to thrombin as
The activation mixtures were then used to titrate 150 nM DAPA in 20 mM Tris, 150 mM NaC1, p H 7.4, a t 22 "C by measuring DAPA fluorescence emission intensity using X,, = 280 nm and X , , = 515-550 nm with a 515 nm cut-off filter. The data were analyzed by the equation (Nesheim et al., 1979;Hibbard et al., 1982) in which [PI,, and [I,],, are the nominal concentrations of protein and ligand, respectively; f and b are the fraction of total ligand concentration free and bound, respectively; n is the number of independent, noninteracting ligand-binding sites/protein molecule, and Kd is the dissociation constant.

FPLC and SDS-PAGE Behaviors of Native and Recombinant Prothrombins -Recombinant prothrombins
isolated after barium precipitation and ion-exchange FPLC were nearly homogeneous as judged by SDS-PAGE (data not shown). Recombinant prothrombins migrated the same as native prothrombin on SDS-PAGE under both nonreducing and reducing conditions (Fig. 2). When eluted with a NaCl gradient from a Mono Q column, wild-type recombinant prothrombin was eluted a t 21.14 min and mutant recombinant prothrombin (Ser"'"+Ala) was eluted a t 21.01 min as compared to 21.26 min for native bovine prothrombin. After the FPLC purification, the yields of wild-type and mutant prothrombin were 0.6-0.8 and 1.5-2.0 pg/ml of cell medium, respectively.

Recombinant Prothrombins Possess Normal Gla Content-
The relative y-carboxyglutamic acid (Gla) contents of the fragment 1 peptides isolated from recombinant and native prothrombins were determined by the Gla-specific [14C]formaldehyde-morpholine method (Wright et al., 1986). This method specifically labels Gla residues with ['4C]formaldehyde. The results are presented in Table I, where the specific "C radioactivity of each modified fragment 1 was obtained using the molecular weight of 22,000 (Mann, 1976). Using the ratio of the specific radioactivity of recombinant prothrombin fragment 1 to that of the native molecule and a value of 10 Gla residues/molecule (Nelsestuen et al., 1974) for native prothrombin fragment 1, the relative values of Gla content of recombinant prothrombin fragment 1 were calculated (Jorgensen et al., 1987). Mutant recombinant prothrombin con-  " 10 Gla residues/native prothrombin fragment 1 was taken as the standard to obtain relative values for the recombinant molecules.
, , m e (r,"n) FIG. 3. Enzymatic activities of prothrombin activation products toward S2238 and fibrinogen. 1.2 PM native or recombinant prothrombins were activated with 0.5 nM factor X*, 5 nM factor v , 10 p M Ps/Pc (25/75) suv, and 5 mM CaCh in 20 mM Tris, 150 mM NaC1, pH 7.4, at 37 "C for 45 min. An aliquot of each incubation was taken to test the activity toward synthetic chromogenic substrate S2238 ( A ) and fibrinogen ( B ) . Data are shown in both figures for native bovine prothrombin (circles), wild-type recombinant prothrombin (triangles), mutant recombinant prothrombin (squares), and the control incubation without prothrombin (inverted triangles). tained 9.9 & 0.5 mol of Gla residues/mol of fragment 1, and wild-type recombinant prothrombin contained 9.4 f 0.4 mol of Gla residues/mol of fragment 1. Fig. 2, activation of wild-type recombinant prothrombin by the full prothrombinase complex was similar both in rate and extent of activation to the activation of native prothrombin. By contrast, the activation of the mutant prothrombin (Ser"8+Ala) was much slower than that of the native molecule under the same conditions. The initial rate of recombinant prothrombin activation relative to that of native prothrombin, as quantitated with the scanning densitometer, was 97% for wild-type prothrombin and 28% for mutant prothrombin.

Mutant Thrombin Showed No Activity Toward S2238 and
Fibrinogen-Thrombin derived either from wild-type prothrombin or mutant prothrombin by the prothrombinase activation was tested for its activities toward synthetic chromogenic substrate S2238 and fibrinogen and compared to native bovine thrombin (Fig. 3). As expected, wild-type thrombin had the same enzymatic activities as native thrombin, while the mutant thrombin (SerS2'+Ala) displayed no activity toward these substrates.
The reaction mixtures of the recombinant prothrombins activated by Taipan snake venom, in which thrombin was the activation product as judged by SDS-PAGE, were also tested. The same results (data not shown) were obtained as were obtained with the prothrombinase-derived thrombin.

SDS-PAGE Analysis of Recombinant Prothrombin Actiuation by
Ecarin-When activated with ecarin, wild-type recombinant prothrombin behaved the same as native prothrombin, as analyzed with SDS-PAGE (Fig. 4). Activation of the wildtype prothrombin was almost completed a t 90 min, and most prothrombin was converted to meizothrombin des fragment 1, which is the product of bovine meizothrombin autolysis (Rosing and Tans, 1988). Activation of mutant prothrombin was also completed by 90 min as judged by the reducing gel ( Fig. 4 B ) , but the product was meizothrombin instead of meizothrombin des fragment 1, showing mutant meizothrombin was not subject to autolysis as were the native and wildtype molecules.

DAPA Assay of Mutant Prothrombin Activation by Ecarin or Taipan
Venom-When monitored with the DAPA fluorescence assay, the emission intensity of the mutant prothrombin reaction mixture activated by ecarin (line 2, Fig. 5 A ) reached a plateau after 300 s. The native prothrombin reaction mixture (line 1, Fig. 5A) also reached a maximum a t about 300 s and subsequently declined. The later behavior is characteristic of formation of meizothrombin followed by its subsequent autolysis (Hibbard et al., 1982;Doyle and Mann, 1990), as shown in the SDS-PAGE results in Fig. 4. Also consistent with Fig. 4, the initially formed mutant meizothrombin was stable (line 2, Fig. 5A).
The most interesting aspect of the data shown in Fig. 5 is that the fluorescence intensity of the mutant prothrombin activation mixture was only 33% of that of native molecule at the maximum when activated by ecarin. In addition, the fluorescence intensity after activation of mutant prothrombin to thrombin with Taipan venom was only 40% of the fluores- cence intensity observed for the native thrombin-DAPA complex. This could be due either to reduced binding of DAPA or to a reduced quantum yield of the DAPA-mutant active site complex as compared to the DAPA-native active site complex. The DAPA fluorescence emission spectrum of the mutant prothrombin reaction mixture activated by ecarin was similar (X, , , = 525 nm) to that of the native activation mixture, measured a t about 300 s after initiation of the activation (data not shown). The DAPA fluorescence emission spectrum of the mutant prothrombin reaction mixture activated by Taipan snake venom was also similar (Xmax = 540 nm) to that of the native activation mixture (data not shown).
Mutated Thrombin Active Site Showed a Weaker DAPA Binding Affinity-As titrated with the same amount of DAPA under the same conditions, the fluorescence intensity of the complex of DAPA and mutant thrombin was 74% of that of the complex of DAPA and native thrombin when both titrations reached saturation (data not shown). Analysis of the titration data using Equation 1 gave a stoichiometry n = 1.22 mol of DAPA/mol of mutant thrombin and a dissociation constant Kd = 5.4 X M (Fig. 6 B ) as contrasted to n = 1.02 mol of DAPA/mol of native thrombin and K d = 1.7 x lo-* M (Fig. 6 A ) . By comparison, n = 1.20 & 0.14 mol of DAPA/mol of bovine thrombin and K d = (1.9 & 0.6) X M were reported by Hibbard et al. (1982).

DISCUSSION
The catalytic triad residues in thrombin, as for other members of the serine protease family, are identified as His''Gfi,  (Nesheim et al., 1979;Hibbard et al., 1982). and Ser' " (Mann et al., 1981). Because of the essential function of thrombin in blood coagulation, no natural prothrombin mutated in any of these 3 catalytic residues is likely to be found. In the current study, a mutant bovine prothrombin (Ser"""+Ala) cDNA has been constructed, subcloned, and expressed in CHO. The expressed mutant recombinant prothrombin, as compared to native prothrombin, is fully ycarboxylated in CHO cells in the presence of vitamin K at a yield of 1.5-2.0 pg/ml. The mutant prothrombin (Ser5"-Ala), with the same mobility as native prothrombin on an anion exchange FPLC column and on SDS-PAGE, can be activated to meizothrombin and thrombin as are native and wild-type expressedprothrombin. The mutant meizothrombin obtained by ecarin activation is very stable and does not undergo autolysis even in the absence of thrombin active site inhibitors. Thus, it provides a useful tool to examine the role of meizothrombin as an important intermediate in prothrombin activation by the full prothrombinase. The availability of a stable form of meizothrombin will allow determination of the secondary or tertiary structure of meizothrombin as well as of its interactions with other coagulation factors and with the phospholipid surface. The mutant thrombin, as expected, does not cleave fibrinogen and does not hydrolyze thrombinspecific chromogenic substrates. This inactive mutant thrombin is suitable for studying the binding interactions of thrombin with its substrate and with inhibitors in the absence of catalysis.

An Active
The successful expression of the mutant prothrombin in vitro might also benefit crystallographic studies of prothrombin, thrombin, and even meizothrombin. So far, only the crystallographic structures of prothrombin fragment 1 and of thrombin inhibited with D-Phe-Pro-Arg chloromethylketone or hirudin have been reported (Tulinsky et al., 1988;Skrzypczak-Jankun et al., 1989;Bode et al., 1989;Rydel et al., 1990;Gliitter et al., 1990). DAPA, a specific inhibitor of the thrombin active site (& = 1.7 X lo-' M), does not bind to prothrombin or to prethrombin 1, in which both Arg274-Thr275 and Arg323-Ile324 remain intact, while it binds at a lower affinity ( K d = 5.9 X