Examining Thrombin Hydrolysis of the Factor XIII Activation Peptide Segment Leads to a Proposal for Explaining the Cardioprotective Effects Observed with the Factor XIII V34L Mutation*

In the blood coagulation cascade, thrombin cleaves fibrinopeptides A and B from fibrinogen revealing sites for fibrin polymerization that lead to insoluble clot formation. Factor XIII stabilizes this clot by catalyzing the formation of intermolecular cross-links in the fibrin network. Thrombin activates the Factor XIII a2 dimer by cleaving the Factor XIII activation peptide segment at the Arg37-Gly38 peptide bond. Using a high performance liquid chromatography assay, the kinetic constants K m , k cat, andk cat/K m were determined for thrombin hydrolysis of fibrinogen Aα-(7–20), Factor XIII activation peptide-(28–41), and Factor XIII activation peptide-(28–41) with a Val34 to Leu substitution. This Val to Leu mutation has been correlated with protection from myocardial infarction. In the absence of fibrin, the Factor XIII activation peptide-(28–41) exhibits a 10-fold lower k cat/K m value than fibrinogen Aα-(7–20). With the Factor XIII V34L mutation, decreases in K m and increases ink cat produce a 6-fold increase ink cat/K m relative to the wild-type Factor XIII sequence. A review of the x-ray crystal structures of known substrates and inhibitors of thrombin leads to a hypothesis that the new Leu generates a peptide with more extensive interactions with the surface of thrombin. As a result, the Factor XIII V34L is proposed to be susceptible to wasteful conversion of zymogen to activated enzyme. Premature depletion may provide cardioprotective effects.

Gly 15 peptide bond and fibrinopeptide B (FpB) is released. Removal of the fibrinopeptides leads to exposure of fibrin polymerization sites that react to form an insoluble blood clot (reviewed in Ref. 1).
Activated Factor XIII helps stabilize this clot structure by catalyzing the formation of intermolecular ␥-glutamyl-⑀-lysine cross-links in the fibrin network and in fibrin-enzyme complexes. Factor XIII is a member of a family of enzymes known as transglutaminases that have a catalytic triad, similar to cysteine proteases, composed of amino acids Cys 314 , His 373 , and Asp 396 . In plasma, Factor XIII is expressed as a zymogen of the form a 2 b 2 . In the presence of thrombin and calcium, the a 2 unit is released and activated. By contrast, platelet Factor XIII is expressed as the zymogen a 2 unit (reviewed in Ref. 2).
The Factor XIII a 2 dimer contains in the N-terminal portion of each monomer a sequence known as the activation peptide (3,4). Each activation peptide segment crosses the dimer interface and extends over the catalytic site of the opposing Factor XIII a subunit. Cleavage of the activation peptide segments by thrombin at the Arg 37 -Gly 38 peptide bond aids in exposure of the Factor XIII active site residues. X-ray crystallography on thrombin-cleaved Factor XIII indicates that the activation peptides do not immediately dissociate following cleavage (4). The activation of plasma Factor XIII a 2 b 2 by thrombin is accelerated in the presence of fibrin-I (5-7) resulting in an 80-fold increase in the k cat /K m value for this reaction. A segment of fibrin-I is known to bind at thrombin anionexosite I, whereas other separate segments of fibrin-I are hypothesized to bind to Factor XIII.
A mutation within the Factor XIII AP segment, Val 34 to a Leu (V34L), has been correlated with protection against myocardial infarction (15)(16)(17)(18). This protection may be mediated through the formation of weaker fibrin structures. In contrast, a tendency to form rigid fibrin networks has been reported in men who have experienced an episode of myocardial infarction at an early age (19). In addition to its possible role in minimizing myocardial infarction, the Factor XIII V34L mutation has also been correlated with protection against venous thromboembolism (20,21) and increased susceptibility to cerebrovascular disease (22).
Factor XIII V34L is a common polymorphism estimated to occur in 20 -25% of the Caucasian population (23)(24)(25). The characteristics and prevalence of Factor XIII V34L have been examined in different ethnic groups including Finnish, Russian, German, Japanese, and Native American Pima (23)(24)(25). A G to T point mutation in codon 34, exon 2 of the Factor XIII a subunit gene, is responsible for the Val (GTG) to Leu (TTG) change. The beneficial effects of the V34L mutation are negated by the presence of additional hemostatic risk factors such as increased plasminogen activator inhibitor-1, decreased plasmin, and increased tissue plasminogen activator (17). These agents compete for ability to maintain or lyse the fibrin clots. Possible explanations for how V34L protects against myocardial infarction include the following: 1) the mutation hinders activation of Factor XIII a 2 thus slowing down initiation of fibrin cross-linking, or 2) the mutation accelerates activation of the Factor XIII a 2 before sufficient fibrin is produced and as a result leads to early depletion of usable levels of Factor XIII.
To aid in elucidating the source of the cardioprotective effects of Factor XIII V34L, a kinetic study was initiated. An HPLC assay was used to compare the ability of thrombin to hydrolyze a segment of the Fibrinogen A␣ chain (Fbg A␣-(7-20)), the native Factor XIII Activation Peptide (FXIII AP-(28 -41)), and the mutant peptide (FXIII AP- (28 -41), V34L). The kinetic values K m , k cat , and k cat /K m were determined for each peptide. The Fbg A␣ amino acids 7-20 bind along the active site surface of the enzyme thrombin and serve as a good model for the complete Fibrinogen A␣ chain, a major substrate of this serine protease. Kinetic results indicate that, in the absence of fibrin, hydrolysis of the native FXIII AP-(28 -41) has a lower k cat /K m value than hydrolysis of Fbg A␣- (7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20). The Factor XIII V34L mutation yields a peptide with improved kinetic properties that begin to approach those of Fbg A␣- (7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20). Activation of Factor XIII V34L by thrombin is proposed to be accelerated over that of wild-type Factor XIII. A review of kinetic properties and the x-ray crystal structures of known substrates and inhibitors of thrombin leads to the hypothesis that the Val to Leu substitution generates a peptide exhibiting more extensive interactions with the surface of thrombin. As a result of these characteristics, the mutant Factor XIII may lead to premature depletion of transglutaminase activity and thus to decreased levels of crosslinked products.

EXPERIMENTAL PROCEDURES
Materials-Bovine plasma barium citrate eluate, bovine Fibrinogen, and Echis carinatus snake venom were purchased from Sigma. HPLC grade acetonitrile, trifluoroacetic acid, Trizma (Tris base), NaCl, PEG 8000, and phosphoric acid were purchased from Fisher and Sigma.
Synthetic Peptides-Peptides based upon residues 7-20 of the human Fibrinogen A␣ chain were synthesized by solid-phase methodologies by the Cornell University Biotechnology Resource Center (Ithaca, NY). Peptides based on residues 28 -41 of the human Factor XIII activation Thrombin Preparation-Bovine thrombin was isolated and purified by methods similar to those of Ni et al. (12) and Maurer et al. (26). Briefly, bovine plasma barium citrate eluate containing prothrombin was activated with E. carinatus snake venom. The resultant thrombincontaining mixture was desalted on Amersham Pharmacia Biotech PD-10 columns and then purified on a Amersham Pharmacia Biotech Mono S cation exchange column (HR 5/5) using a Waters HPLC system (model 600 controller and pump system, Rheodyne manual injector, and a 2487 dual absorbance detector). The pooled thrombin solution was subjected to ultrafiltration. The final concentration of protein was determined by using an extinction coefficient, E 1% ϭ 19.5 at 280 nm (27).
For this project, bovine thrombin was used as the enzyme and the synthetic substrate peptides were based on human sequences. These complexes were prepared so that a direct comparison could be made with previously published HPLC kinetic results (28 -30) and NMR/xray studies (11)(12)(13)(14) on thrombin-ligand complexes. Thrombin exhibits a high conservation of sequence between human and bovine forms (31). There are no differences in the residues involving the active site, the thrombin ␤-insertion loop (also called the Trp 60D loop), or the allosteric Na ϩ -binding site. The minor differences that do exist between species are not anticipated to interfere with interactions of the substrate peptides at the thrombin active site surface (32).
Peak areas corresponding to the hydrolyzed products were converted into concentrations by applying a linear calibration curve. These calibration curves were generated from HPLC runs of the products Fbg A␣-(7-16), FXIII AP- (28 -37), and FXIII AP-(28 -37) V34L. Stock concentrations of these peptides were determined by quantitative amino acid analysis by the Cornell Biotechnology Resource Center. Standard peptide calibration curves over the range of 8 to 120 M were generated. The same HPLC gradients used for the kinetic assays were employed here.
Initial velocities (in micromolar/s) for the thrombin-catalyzed reactions were determined for each peptide concentration from the slopes of product
Kinetics Evaluation of Thrombin Hydrolysis of Fibrinogen A␣-like and Factor XIII-like Substrates-Nonlinear regression analysis values for K m , k cat , and k cat /K m are shown in Table II.  (Table II). The substitution of leucine for valine at position 34 has led to the creation of a substrate with improved kinetic properties over those of the wild-type FXIII AP segment. The properties begin to approach those of Fbg A␣- (7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20). Taking the error limits into account, the K m values for FXIII AP-(28 -41) V34L versus Fbg A␣- (7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20) are quite similar (272 Ϯ 57 versus 312 Ϯ 42 M). In the presence of the V34L mutation, the k cat value increased for Factor XIII activation peptide cleavage from 6.4 Ϯ 0.03 to 18.5 Ϯ 1.6 s Ϫ1 and is now approximately half that of Fbg A␣-(7-20)  5,7,9,11, and 15 min. Each quenched time point was run on a Brownlee Aquapore C8 column using a Waters HPLC system ϩ autosampler. A linear gradient of 15% acetonitrile, 0.09% trifluoroacetic acid in water to 50% acetonitrile, 0.09% trifluoroacetic acid in water in 20 min at a flow rate of 1 ml/min was employed. The original substrate peptide FXIII AP- (28 -41) eluted at approximately 9.3 min whereas the hydrolyzed product FXIII AP- (28 -37) eluted at approximately 4.5 min. The fragment FXIII AP- (38 -41) was released in the load eluent. (39.3 Ϯ 2.6 s Ϫ1 ). Furthermore, the amount of thrombin used to generate the kinetic results for the V34L work was the same as those used for Fbg A␣- (7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20). A review of the k cat /K m values suggests that thrombin exhibits a greater specificity for the V34L mutant sequence than the wild-type sequence. The results reported here, for bovine thrombin in the presence of human Factor XIII AP segments, are consistent with the findings of Kohler et al. (15) on Factor XIII V34L isolated from patients. Preliminary data from their laboratories indicate that lower thrombin levels are required to activate Factor XIII V34L than to activate wild-type Factor XIII. DISCUSSION The kinetics associated with thrombin hydrolysis of a Fibrinogen A␣-like peptide and two Factor XIII activation peptide segments have been examined. A Val to Leu substitution at position 34 of the Factor XIII activation peptide segment generates a substrate with improved kinetic properties relative to that of the wild-type sequence. A review of the sequence, kinetic, and structural properties of thrombin substrates/inhibitors leads to a proposal for how Factor XIII V34L might promote beneficial cardiovascular effects.
Analysis of Kinetic Results and Substrate Peptide Amino Acid Sequence-The peptide substrates analyzed here contain amino acids that only interact with thrombin along a surface surrounding the active site of the enzyme. Work with peptides of this length allows for examination of the influence of individual substrate amino acids on the catalytic properties of thrombin. The human Factor XIII activation peptide segment contains 34 VVPR 37 at the P 4 -P 1 substrate positions. The strongest benefits for thrombin binding are likely derived from the Pro and Arg residues (33)(34)(35)(36). X-ray crystallographic studies (9, 37) on PPACK (D-Phe-Pro-Arg-chloromethylketone) bound to thrombin have shown that the Pro at P 2 is positioned to interact with a hydrophobic cage formed by the thrombin side chains of Trp 215 , Leu 99 , His 57 , Tyr 60A , and Trp 60D . This region is also known as the apolar-binding site.
Thrombin exhibits greater substrate specificity for the Factor XIII activation peptide segment when the 34 VVPR 37 sequence is converted to 34 LVPR 37 . The presence of the larger hydrocarbon side chain on leucine may promote further interactions with the thrombin surface, allowing for more contact within the apolar-binding site on thrombin. The new leucine may promote the decreases in K m relative to the wild-type sequence and, as a result of better interactions with the thrombin surface, the increase in k cat . Together these changes have allowed the k cat /K m value to increase 6-fold relative to the wild-type sequence.
The Factor XIII V34L mutant sequence 34 LVPR 37 is similar to the 38 LDPR 41 of the PAR1 thrombin receptor. However, hydrolysis of a thrombin receptor PAR1 peptide ( 29 PESKAT-NATLDPRSFLL 45 ) exhibits a higher K m value than both Factor XIII activation peptides (see Table II) (38 -40). NMR (41) and x-ray (10) studies suggest the LDPR segment binds along the apolar-binding site but the Asp 39 causes electrostatic interference that hinders optimal binding along this thrombin surface (10,37). The Factor XIII AP segment does not contain an interfering acidic residue, and as a result, FXIII-like peptides targeted to the active site of thrombin serve as better substrates than similar PAR1-like peptides.
In contrast to the FXIII-and PAR1-based peptides, the Fbg A␣-(7-20) peptide does not generate beneficial interactions with the surface of thrombin through a proline. In the bound structure, a multiple-turn conformation involving Fbg A␣ Phe 8 , Leu 9 , and Val 15 provides key sites for interaction with the thrombin apolar-binding site. As a result of this interaction, the non-optimal nature of the A␣ chain can be overcome and the sequence can be cleaved effectively. The k cat /K m value for hydrolysis of the Factor XIII AP segment only begins to approach that of Fbg A␣- (7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20) in the presence of the V34L substitution.
The kinetic results obtained with the Factor XIII activation peptides thus provide convincing evidence that a relatively conservative amino acid substitution to the P 3 substrate position causes a significant improvement in substrate specificity. An increase in the k cat value for hydrolysis of the mutant Factor XIII sequence plays a major role in this change. The different kinetic parameters obtained for hydrolysis of the peptide models therefore help explain the reports that human Factor XIII V34L requires less thrombin for activation than the wild-type Factor XIII (15).
Proposed Model to Explain the in Vivo Effects of the Factor XIII V34L Mutation-In nature, the proteins thrombin, fibrin, and plasma Factor XIII function in a triad, with fibrin serving as a cofactor for Factor XIII activation (5,6). Recent results from Brummel et al. (7) on the extrinsic coagulation pathway of human whole blood suggest that Factor XIII activation and its subsequent cross-linking capabilities occur immediately in response to removal of fibrinopeptide A from fibrinogen. In the presence of the resultant fibrin-I, activation of plasma Factor XIII is accelerated with the k cat /K m for a ternary complex of thrombin, Factor XIII, and fibrin becoming comparable to the k cat /K m for thrombin-fibrinogen interactions. The use of a triad in the enzymatic blood coagulation mechanisms ensures that activated plasma Factor XIII is generated close to its substrates, the linear fibrin monomers, thereby minimizing wasteful generation of the Factor XIII when fibrin levels are low. This sort of strategy allows Factor XIII zymogen levels to remain conserved when substrate is not available (42,43).
Information from the blood coagulation/anticoagulation cascades may be used to propose why the Factor XIII V34L mutation may exert cardioprotective effects. Individuals bearing this mutation may be capable of producing higher levels of activated plasma Factor XIII than those bearing the wild-type Factor XIII. Furthermore, the Factor XIII V34L may have less of a dependence on fibrin-I to accelerate its activation by thrombin. Thus, the presence of the V34L mutation may lead to wasteful generation of activated Factor XIII. The excess activated Factor XIII a 2 subunits may then be susceptible to further proteolysis by thrombin at the secondary Factor XIII cleavage site Lys 513 -Ser 514 . Proteolytic enzymes other than thrombin could also be involved in inactivation or degradation processes. As a consequence of the V34L mutation, the ability to promote extensive covalent cross-linking of fibrin and fibrinenzyme complexes may become diminished prematurely (42)(43)(44)(45).
This type of premature depletion of an activated enzyme has also been observed for the conversion of plasminogen to plasmin (46). Plasmin is involved in promoting lysis of the blood clots. If tissue plasminogen activator is employed in the zymogen activation process, a fibrin surface is required, and plasmin is produced in the vicinity of its own substrate (47). By contrast, if streptokinase is employed, the fibrin surface is no longer required (48,49). Plasmin can be generated at a site distant from fibrin, and the levels of plasmin may become prematurely depleted before this plasmin has an opportunity to reach its target.
In addition to playing a role in Factor XIII activation, the V34L mutation has also been reported to affect the crosslinking activity of this transglutaminase. Work with native and recombinant Factor XIII indicates that the V34L enzyme exhibits a higher specific activity than wild-type enzyme (50,51). This result suggests that the activation peptide may actually participate in promoting the transglutaminase reaction of Fac-tor XIII. Somehow this participation is further amplified and/or changed in the presence of the mutant sequence. The activation peptide does not leave immediately following cleavage and is proposed to remain bound until after introduction of Factor XIII substrates (3,4). In the presence of the V34L mutation, the number and/or type of substrates that can be accommodated by the Factor XIII active site surface and subjected to cross-linking may become altered. Further work on binding of substrates to the intact enzyme is necessary to elucidate the source of the increase in specific activity for the mutant and to correlate this increase with the observed physiological responses. The increase in Factor XIII specific activity may contribute a yet unidentified role in protection against myocardial infarction. Alternatively, the mutation may generate an increased catalytic effect that becomes diminished before its benefits can be realized.
Conclusions-Our kinetic studies reveal that thrombin-catalyzed hydrolysis of a Factor XIII-based activation peptide containing the V34L mutation leads to decreases in K m , increases in k cat , and as a result, increases in k cat /K m relative to the wild-type sequence. The Factor XIII sequence VVPR is converted to LVPR. The Leu substitution is proposed to make more contacts with the apolar-binding site region on thrombin eventually leading to increases in substrate specificity relative to the wild-type sequence. As a result of these changes, Factor XIII V34L is hypothesized to be susceptible to wasteful conversion of zymogen to activated enzyme. Premature depletion may provide protective effects against cardiovascular disease.