Catalytic Life of Activated Factor XIII in Thrombi
Implications for Fibrinolytic Resistance and Thrombus Aging
Background—Because the increased fibrinolytic resistance of older thrombi may be caused by the continuous cross-linking action of fibrin-bound activated factor XIII (FXIIIa), we examined the persistence of FXIIIa catalytic activity in clots of various ages.
Methods and Results—The time-related changes in FXIIIa activity in clots was measured with (1) α2-antiplasmin (α2AP), a physiological glutamine substrate; (2) α2AP13–24, a peptide; and (3) pentylamine, a nonspecific lysine substrate. The cross-linking of α2AP, α2AP13–24, and pentylamine into fibrin by clot-bound FXIIIa declined rapidly with half-lives of 19, 21, and 26 minutes, respectively. Mutational studies showed that glutamine 14 (but not glutamine 3 or 16) and valine 17 of α2AP13–24 were required for efficient cross-linking to fibrin. The loss of activity was not due primarily to FXIIIa proteolysis and was partially restored by reducing agents, suggesting that oxidation contributes to the loss of the enzyme’s activity in clots. In vivo, the ability of thrombus-bound FXIIIa to cross-link an infused α2AP13–24 peptide into existing pulmonary emboli also declined significantly over time.
Conclusions—FXIIIa cross-links α2AP and an α2AP peptide, in a sequence-specific manner, into formed clots with a catalytic half-life of ≈20 minutes. This indicates that FXIIIa activity is a hallmark of new thrombi and that the antifibrinolytic cross-linking effects of FXIIIa are achieved more rapidly in thrombi than previously believed.
The successful dissolution or fibrinolysis of blood clots has life-saving effects in patients with thrombotic disease. Yet despite improvements in the design and administration of plasminogen activators, there is evidence that all types of vascular thrombi resist fibrinolysis and that this resistance increases with thrombus age.1 2 Coronary thrombi (>4 hours old) are more resistant to the lytic effects of plasminogen activators, and the success of thrombolysis in venous thromboembolism diminishes significantly with the clinical age of thrombi.3 4 The mechanism(s) responsible for enhanced fibrinolytic resistance of thrombi with age remain unknown.
Experimental studies of pulmonary thromboembolism show that activated coagulation factor XIII (FXIIIa) makes new thrombi resistant to lysis by physiological and pharmacological plasminogen activators.5 A transglutaminase, FXIIIa cross-links proteins in the fibrin thrombus by forming an amide bond between specific glutamine and lysine residues, which produces a mechanically stronger clot.6 7 8 9 Fibrinolytic resistance is achieved when FXIIIa covalently links α2-antiplasmin (α2AP), the plasmin inhibitor, to fibrin.10 Cross-linking of fibrin α-chains and/or γ-chains by FXIIIa also increases the resistance to plasmin.11 In vitro, FXIIIa-mediated cross-linking is initiated at the time of clotting and appears to be progressive over the course of minutes to hours. Although there are no direct studies of FXIIIa cross-linking rates in human thrombi, studies in rabbits suggest that FXIIIa cross-linking of α-chains may continue for ≥6 hours after clotting.12 This persistence of FXIIIa cross-linking seemed plausible because there were no known physiological inhibitors of FXIIIa in humans and the regulation of the catalytic half-life of the enzyme was poorly understood. Because more extensively cross-linked thrombi are less sensitive to degradation by plasmin, the increased resistance of older thrombi to fibrinolysis could be mediated, in part, by FXIIIa.
In this study, we examined the catalytic activity of FXIIIa in vitro and in vivo after clotting. These findings have implications for the potential therapeutic use of FXIIIa inhibitors and for diagnostic or treatment strategies that seek to target newly formed thrombi.
Materials were obtained from the following suppliers: 5-(biotinamido)pentylamine, Pierce; α2AP, Calbiochem; biotinylated α2AP peptides (high-performance liquid chromatography–purified >95%), Advanced Chemtech; bovine thrombin, DTT, and tissue transglutaminase from guinea pig liver, Sigma; calcium chloride, Mallinckrodt; purified FXIII, human fibrinogen, plasmin, and rabbit anti-human fibrin antibody, American Diagnostica; fresh-frozen human plasma pooled from random donors, MGH Blood Bank; mouse antifibrinogen antibody 4A5 and mouse anti-FXIII antibody 309, Brscom; iodoacetamide, Calbiochem; [125I]NaI, NEN-Dupont; and the histological preparations avidin-biotin complex (ABC) and 3,3′-diaminobenzidine (DAB), Vector Laboratories.
Ferrets (0.9 to 1.0 kg) were obtained from Marshall Farms, North Rose, NY; mice (Black C57, 4 to 8 weeks old), Jackson Laboratories, Bar Harbor, Me; ketamine, Fort Dodge Laboratories; acepromazine maleate, Fermenta Animal Health Co; surgical instruments, VWR and Roboz; Bard Parker surgical blades and polyethylene tubing, Becton Dickinson; silk suture, American Cyanamid Co; Surflo IV catheter, 20 gauge, Terumo Medical Corp; and 3-way stopcocks and micro tissue grinder, VWR.
Assays of FXIIIa Catalytic Life and Substrate Specificity
The persistence of FXIIIa cross-linking activity in clots was examined with various substrates: human 125I-labeled α2AP, 5-(biotinamido)pentylamine, and a peptide, α2AP13–24 (in which lysine 24 was biotinylated). Clots were prepared (50 μL) in duplicate by combining fibrinogen (2 mg/mL final), CaCl2 (2 mmol/L final), buffer (Tris-buffered saline, pH 7.4), and thrombin (1 U/mL) and incubating at 37°C. The FXIIIa substrates 5-(biotinamido)pentylamine (0.5 mmol/L final), 125I-α2AP (70 μg/mL final), and α2AP13–24 (0 to 1 mmol/L final) were added 0 to 240 minutes after clotting with thrombin. In 1 set of control samples, FXIIIa was inhibited with iodoacetamide (10 mg/mL) before the addition of substrate; in another set, no substrate was added. After 2 hours of incubation at 37°C, iodoacetamide was added to stop the reaction. Clots were then centrifuged at 14 000 rpm for 2 minutes, washed and compressed in 1 mL saline, and centrifuged again. The washed clots were solubilized in 100 μL of 9 mol/L urea, pH 9.0, at 37°C for 60 minutes. Then 100 μL of SDS-reducing sample buffer was added, and samples were held at 85°C for 30 minutes and examined by SDS-PAGE. Samples containing cross-linked α2AP13–24 and pentylamine substrates were electroblotted to polyvinylidine difluoride (PVDF) membranes and detected by 125I-streptavidin followed by phosphorimaging or by streptavidin-peroxidase and the developing agents 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium (NBT-BCIP) as we have described.5 Phosphorimages were examined under conditions in which there was a linear relationship between substrate incorporation and pixel volume. In experiments examining the effect of specific amino acid residues on the cross-linking of the α2AP13–24 peptide to fibrin, cross-linking was detected by immunoblotting with an anti-peptide antibody directed against this epitope followed by 125I-labeled protein A and phosphorimaging as described.13 Plasma clots were formed by mixing 25 μL pooled human fresh-frozen plasma (sodium citrate) with 25 μL 30 mmol/L CaCl2 instead of human fibrinogen.
To examine the potential degradation of FXIIIa in clots, samples from these experiments were immunoblotted with a monoclonal antibody directed against the A subunit of FXIII as described.14
Cross-Linking of α2AP13–24 Into Fibrin by FXIIIa and Tissue Transglutaminase
Fibrin clots (50 μL) were prepared in duplicate by mixing FXIII-free fibrinogen (2 mg/mL final), α2AP13–24 (0.5 mmol/L final), CaCl2 (2 mmol/L final), human plasma FXIII (100 nmol/L final), and guinea pig tissue transglutaminase (100 nmol/L final) and thrombin (1 U/mL final) and incubating at 37°C for 2 hours. The cross-linking of α2AP13–24 was detected as described above.
Cross-Linking of α2AP13–24 Into Pulmonary Emboli In Vivo
We have previously used a pulmonary embolism model in male ferrets (≈1.0 kg).15 All in vivo experiments were approved by the Harvard Medical Area Standing Committee on Animals. Clots were formed with human fresh-frozen plasma (45 μL) mixed with trace amounts of 125I-labeled fibrinogen (≈100 000 cpm/clot), 2.5 μL of 0.4 mol/L CaCl2, and 2.5 μL thrombin (1000 U/mL) for 20 minutes at 37°C. After 3 washes, 6 clots were embolized into 2 animals in each of 3 groups: 2 groups had normal clots, and 1 group received clots washed in iodoacetamide (10 mg/mL) and EDTA (10 mmol/L final) to inhibit clot-associated FXIIIa. After embolization, α2AP13–24 (final concentration of 0.5 mmol/L) was infused (except in control animals). Four hours after embolization, the ferrets were euthanized by CO2 inhalation, and the pulmonary emboli were isolated.
Samples were weighed and homogenized in a 200-μL solution (24 mmol/L Tris Base, 476 mmol/L Tris HCl, 50 mmol/L MgCl2, 1 mg/mL DNAse I, and 0.25 mg/mL RNAse A), incubated on ice for 10 minutes, and centrifuged at 12 000 rpm at 4°C for 20 minutes. The pellet was dissolved in 200 μL of sample buffer, heated until dissolved, and analyzed by SDS-PAGE as described above.
Mouse Pulmonary Embolism Experiments
Clots were formed with 22.5 μL of whole mouse blood in polyethylene tubing after mixing with trace amounts of 125I-labeled fibrinogen, 1.25 μL of 0.4 mol/L CaCl2, and 1.25 μL thrombin (1000 U/mL) at 37°C for 15 and 30 minutes. The right and left internal jugular veins were exposed in anesthetized mice. Before embolization, 100 U/kg of heparin and 1.5 mg/kg of P-PACK were injected intravenously to inhibit endogenous thrombin activity. One vessel was used to embolize the clot, and the other was used for infusion of α2AP13–24 immediately after embolization. Two hours after embolization, the thrombi were isolated and subjected to SDS-PAGE as described above.
Lung tissue was immersed in 30% sucrose at 4°C overnight, embedded in OCT, cut into 10-μm sections, and fixed in 100% methanol (5 minutes). Samples were pretreated with 3% H2O2 (20 minutes), washed in PBS 3 times, and treated with 10% normal goat serum (20 minutes). Samples were then incubated with rabbit anti-human fibrin antibody (1:400 dilution, 2.5 μg/mL) at 4°C overnight, washed twice (in high-salt PBS and regular PBS), and incubated with goat anti-rabbit peroxidase-labeled antibody (1:100, 1 hour). After 3 washes in PBS, sections were developed in DAB, counterstained in 0.1% methyl green, and mounted. For the detection of α2AP13–24, sections were developed with the ABC reagent and developed as described above.
Data were analyzed with either an unpaired Student’s t test or a 1-way ANOVA with the Bonferroni-Dunn procedure for testing multiple comparisons.
To examine the catalytic half-life of FXIIIa in thrombi with physiological substrates, we measured the cross-linking of α2AP to fibrin. The cross-linking of 125I-α2AP was greatest when it was added synchronously to fibrinogen with thrombin to initiate clotting (Figure 1⇓). When added to older clots, the amount of α2AP cross-linking dropped off rapidly to the levels of controls in which FXIIIa had been inhibited with iodoacetamide, or in lanes containing no 125I-α2AP. The cross-linking of 125I-α2AP to fibrin declined exponentially with a half-life of 19 minutes (r2=0.92, Figure 1⇓, bottom, semilog plot).
One potential explanation for the decline in the cross-linking of α2AP to fibrin was that the size of α2AP (70 000 Da) inhibited its diffusion into the assembling fibrin meshwork. To examine this possibility, we used a small α2AP peptide (α2AP13–24, ≈1600 Da), which should readily permeate the developing thrombus. We and others have verified that this peptide5 or related peptides16 17 can compete with cross-linking of α2AP to fibrin when added during clotting. This peptide was biotinylated at lysine 24 to permit detection. The peptide cross-linked into fibrin clots as a function of dose (Figure 2⇓). Compared with control clots, the cross-linking of the α2AP13–24 peptide was seen at concentrations as low as 1.3 μmol/L. Cross-linking was half of maximum at 110 μmol/L and saturated at concentrations of 0.33 to 1 mmol/L. A similar dose-response curve was found for the cross-linking of the α2AP13–24 peptide to fibrin in human plasma (data not shown).
Although FXIIIa can cross-link a number of lysine analogues to glutamine sites in macromolecules such as fibrin, little is known about its specificity for cross-linking glutamine-containing substrates (acyl donors), such as α2AP, to fibrin. To examine this, a panel of α2AP peptides (residues 1 to 24) was created with mutations in residues that could represent other potential cross-linking sites (Q7, Q14) or residues that are conserved among different α2APs from different species (V17, L22; Figure 3⇓, top). Compared with clots containing no peptide or the wild-type peptide α2AP1–24, peptides containing mutations Q14A and V17N were not efficiently cross-linked to fibrin (Figure 3⇓, bottom). In contrast, mutations Q7A, Q16A, and L22N were cross-linked into fibrin at rates not markedly different from the wild-type α2AP1–24 peptide. Similar results were seen when these peptides were used as inhibitors of the cross-linking of α2AP13–24 into fibrin: all peptides but Q14A and V17N competed for cross-linking (data not shown). Taken together, these results indicate that α2AP1–24 and α2AP13–24 are specific glutamine substrates for FXIIIa and suggest that Q14 and V17 are necessary for efficient fibrin cross-linking.
Although FXIIIa is the primary transglutaminase found in plasma, other transglutaminases are found in cells that do not require thrombin for activation. We examined whether the α2AP13–24 peptide could be efficiently cross-linked by tissue transglutaminase. In studies with FXIIIa-deficient fibrinogen, there was little, if any, significant cross-linking of α2AP13–24 to fibrin with clotting (Figure 4⇓, top). Addition of purified tissue transglutaminase (100 nmol/L) caused an increase in cross-linking that was not statistically significant (P>0.05). Equivalent amounts of purified exogenous FXIIIa caused significantly more cross-linking of α2AP13–24 to fibrin, which was significantly blocked by the thrombin inhibitor hirudin (P<0.01, Figure 4⇓, bottom). Because FXIIIa is the only thrombin-dependent transglutaminase, this result indicates that FXIIIa was responsible for the cross-linking of the α2AP13–24 peptide in these fibrinogen preparations.
The α2AP13–24 peptide was used to measure the catalytic life of FXIIIa in formed or forming clots. Consistent with experiments performed with 125I-α2AP (Figure 1⇑), the cross-linking of α2AP13–24 to fibrin was greatest when it was present at the initiation of clotting (Figure 5⇓) and fell rapidly over the course of an hour to nearly undetectable levels. The half-life calculated for α2AP13–24 cross-linking was 21 minutes (r2=0.935), which was comparable to that found for the 125I-α2AP (Figure 1⇑). We also found that FXIIIa showed a similar catalytic half-life for a lysine (acyl acceptor) mimic, pentylamine-biotin, an even smaller substrate (330 Da) than α2AP13–24. In these studies, pentylamine-biotin was cross-linked into fibrin clots with a half-life of 26 minutes (r2=0.744), which was consistent with the cross-linking of α2AP and α2AP13–24.
We examined whether proteolysis of FXIIIa by thrombin accounted for the loss of FXIIIa activity in clots (Figure 6⇓). Quantitative immunoblotting showed that after 1 hour, there was an ≈10% decrease in the amount of FXIIIa catalytic subunit A′, and by 4 hours, there was an ≈20% decrease. In contrast, 125I-α2AP–fibrin cross-linking had declined by ≈80% within 1 hour and by >95% within 4 hours (Figure 1⇑). Because proteolysis alone did not explain the loss of FXIIIa activity, we examined whether oxidation of the multiple reduced cysteines of FXIIIa (including its active-site residue) could account for the loss in catalytic activity of the molecule. Under normal circumstances, when pentylamine or α2AP13–24 was added to 2- or 4-hour-old clots, there was little cross-linking of the substrates to fibrin by FXIIIa (Figure 7⇓). The addition of a reducing agent, DTT, restored the cross-linking of pentylamine or α2AP13–24 to fibrin clots, which suggested that reversible oxidation of FXIIIa was responsible for the loss of cross-linking activity.
To confirm the physiological significance of these findings, we examined the ability of FXIIIa to cross-link α2AP13–24 to existing pulmonary emboli in vivo. Three experimental groups were studied: (1) a control group that received no peptide after embolization, (2) a group that received a 0.5 mmol/L dose of α2AP13–24 peptide after embolization, and (3) a group that received a 0.5 mmol/L dose of α2AP13–24 peptide after embolization of clots that were pretreated with iodoacetamide and EDTA to inhibit residual clot-associated FXIIIa activity. Figure 8A⇓ shows a pulmonary embolus (stained with anti-fibrin antibody) that has occluded a pulmonary arteriole. Compared with a control thrombus (8B) from an animal not receiving α2AP13–24, the thrombus from an animal receiving the peptide infusion (8C) showed specific peroxidase staining indicating incorporation of the α2AP13–24 into the thrombus.
To confirm that the peptide was covalently incorporated into fibrin in the thrombus, the lung tissue samples were subjected to SDS-PAGE, blotted with 125I-streptavidin to detect biotinylated α2AP13–24 (Figure 9⇓, top), and stained with Coomassie blue (Figure 9⇓, bottom). Lung tissue from control animals (normal emboli and no α2AP13–24) showed 2 nonspecific avidin-binding bands at ≈70 and ≈120 kDa in the tissues with and without the embolus (Figure 9A⇓, lanes 1 and 2). In animals receiving the α2AP13–24 peptide, there were similar nonspecific staining bands at 70 and 120 kDa in the lung tissue without emboli (Figure 9C⇓, lane 5). However, the lung tissue containing pulmonary emboli showed a new broad band at ≈65 to 70 kDa and additional bands at 60, ≈100, and ≈140 kDa (Figure 9C⇓, lane 6). The intensity of the bands was diminished in the lung tissue from an animal with a pulmonary embolus in which FXIIIa had been inhibited by iodoacetamide and EDTA (Figure 9B⇓, lane 4). Taken together, the studies in Figures 8⇑ and 9⇓ show that α2AP13–24 is cross-linked into preexisting pulmonary embolus in vivo and that cross-linking is attenuated by inhibition of the FXIIIa in the thrombus.
Additional experiments were performed in mice to investigate whether FXIIIa cross-linking ability also declined in formed pulmonary emboli in vivo. Compared with control mice (no peptide infused), there was a significantly greater amount of α2AP13–24 cross-linking in the 15-minute-old emboli. This cross-linking declined significantly in the 30-minute-old emboli (P<0.02, Figure 10⇓).
The finding that FXIIIa cross-linking caused fibrinolytic resistance in new pulmonary emboli suggested that continued FXIIIa activity may contribute to the increase in fibrinolytic resistance of thrombi with age.5 We found that the catalytic ability of FXIIIa to cross-link its physiological substrate (α2AP, 70 000 Da) into fibrin clots declined in an exponential fashion, with a half-life of 19 minutes. Similar half-lives for FXIIIa-mediated cross-linking were also found in much smaller substrates, such as α2AP13–24 (1600 Da), 21 minutes, and pentylamine-biotin (330 Da), 26 minutes. This argued that the decline in FXIIIa activity was not due to a size-related limitation in the permeation of the substrate into the fibrin clot. The decline in catalytic activity of FXIIIa could not be ascribed to proteolysis of the catalytic subunit of FXIIIa (Figure 6⇑) but rather appeared to be related to potential oxidation (Figure 7⇑). This is consistent with the fact that FXIIIa contains a cysteine residue in its active site that may be vulnerable to inhibition by oxidation. Finally, in vivo studies confirmed our in vitro findings by demonstrating that formed pulmonary emboli covalently incorporated α2AP13–24 in a manner that declined significantly with time. This incorporation was due to the enzymatic activity of the thrombus-bound FXIIIa, because it was significantly attenuated in thrombi in which FXIIIa had been inhibited immediately before embolization.
FXIIIa makes thrombi resistant to plasmin by modifying the fibrin matrix through intermolecular cross-links. FXIIIa-mediated fibrin α-chain polymerization5 12 18 and γ-chain multimerization11 contribute to this fibrinolytic resistance. In addition, the cross-linking of the plasmin inhibitor α2AP to fibrin plays a clear role in neutralizing fibrinolysis.5 10 13 In vitro, the cross-linking of α2AP to fibrin and the formation of γ-dimers occur rapidly, within 2 to 5 minutes. Experimental studies of rabbit thrombi formed in vivo indicate that γ-chain dimerization and α-chain polymerization are readily seen within 7 to 9 minutes of thrombus formation, although small increments in α-chain polymerization were detected after 90 to 320 minutes.12 In one of the few studies of the fibrin structure of human thrombi, all thromboemboli examined showed extensive α-chain polymerization (although the ages of these thrombi were not specified).19 Although previous studies have examined the clearance of circulating FXIII zymogen in deficient patients,20 to the best of our knowledge, these data represent the first estimates of the catalytic life of FXIIIa in clots.
Cross-linking by FXIIIa is highly specific, because only a minority of potential glutamines and lysines in proteins are actually substrates.21 The primary structure, charge, and conformation around the respective glutamine residues determine the suitability of a particular glutamine as a substrate for FXIIIa.9 22 23 Still, no consensus sites have been defined to identify potential FXIIIa substrates. Of the 3 potential glutamine sites in the amino-terminus of α2AP13–24, only glutamine 14 is required. In addition, Val17, which is highly conserved in α2AP from other species, also appears to be necessary.
Our limited in vivo studies confirm that FXIIIa can cross-link a specific substrate (α2AP13–24) into formed emboli and that this cross-linking activity declines significantly with the age of the clot. More extensive studies with imaging techniques (nuclear medicine, MRI, etc) will be required to definitively establish, in vivo, the time-related incorporation of specific FXIIIa substrates into thrombi of different ages and to determine whether significant species differences exist.
Recently, a polymorphism was described in the FXIII gene that modifies a patient’s risk of thrombosis.24 It is possible that this or other individual genetic differences in FXIIIa structure or the relative “oxidative” environment of the blood might also modify FXIIIa half-life in vivo. However, the fact that FXIIIa remains catalytically active only in recent thrombi could be exploited to detect or target therapies to “new” thrombi. The ability to distinguish new from old thrombi might significantly improve our selection of patients for fibrinolytic therapy and could lead to the design of agents that are directed against recently formed clots.
This work was supported in part by NIH grants HL-57314 to Dr Reed and HL-58496 to Dr Robinson. The authors are grateful to Dorothy Zhang for expert histology.
- Received October 13, 1999.
- Revision received March 23, 2000.
- Accepted April 4, 2000.
- Copyright © 2000 by American Heart Association
Lincoff AM, Topol EJ, Califf RM, et al, for the Thrombolysis and Angioplasty in Myocardial Infarction Study Group. Significance of a coronary artery with thrombolysis in myocardial infarction grade 2 flow “patency” (outcome in the Thrombolysis and Angioplasty in Myocardial Infarction trials). Am J Cardiol.. 1995;75:871–876.
Hirsch J, Salzman EW, Marder VJ. Treatment of venous thromboembolism. In: Colman RW, Hirsh J, Marder VJ, et al, eds. Hemostasis and Thrombosis: Basic Principles and Clinical Practice. 3rd ed. Philadelphia, Pa: Lippincott; 1994:1346–1366.
Reed GL, Houng AK. The contribution of activated factor XIII to fibrinolytic resistance in experimental pulmonary embolism. Circulation. 1999;99:299–304.
Mellanby J, Pratt CLJ. Calcium and blood coagulation. Proc Roy Soc B. 1939;128:201–213.
Sakata Y, Aoiki N. Crosslinking α2-plasmin inhibitor of fibrin by fibrin-stabilizing factor. J Clin Invest. 1980;65:290–297.
Siebenlist KR, Mosesson MW. Progressive cross-linking of fibrin γ chains increases resistance to fibrinolysis. J Biol Chem. 1994;269:28414–28419.
Butte AN, Houng AK, Jang I-K, et al. α2-Antiplasmin causes thrombi to resist fibrinolysis induced by tissue plasminogen activator in experimental pulmonary embolism. Circulation. 1997;95:1886–1891.
Kimura S, Tamaki T, Aoki N. Acceleration of fibrinolysis by the N-terminal peptide of α2-plasmin inhibitor. Blood. 1985;66:157–160.
Schrode J, Folk JE. Stereochemical aspects of amine substrate attachment to acyl intermediates of transglutaminases: human blood plasma enzyme (activated coagulation factor XIII) and guinea pig liver enzyme. J Biol Chem. 1979;254:653–661.
Fear JD, Miloszewski KJA, Losowsky MS. The half-life of factor XIII in the management of inherited deficiency.
Micanovic R, Procyk R, Lin W. Role of histidine 373 in the catalytic activity of coagulation factor XIII. J Biol Chem. 1994;269:9190–9194.
Gorman JJ, Folk JE. Structural features of glutamine substrates for transglutaminases: specificities of human plasma factor XIIIa and the guinea pig liver enzyme. J Biol Chem. 1981;256:2712–2715.
Lorand L, Lososowski MS, Miloszewski KJM. Human factor XIII: fibrin stabilizing factor. In: Spaet TH, ed. Progress in Hemostasis and Thrombosis. New York, NY: Grune & Stratton; 1980:245–290.
Catto AJ, Kohler HP, Coore J, et al. Association of a common polymorphism in the factor XIII gene with venous thrombosis. Blood. 1999;93:906–908.
Tamaki T, Aioki N. Cross-linking of α2-plasmin inhibitor to fibrin catalysed by activated fibrin-stabilizing factor. J Biol Chem. 1982;14767–14772.