A chimeric streptokinase with unexpected fibrinolytic selectivity.

Chimeric 59D8-SK was designed to confer fibrin-selectivity to streptokinase by fusion of the Fab fragment of anti-fibrin antibody 59D8 to the N-terminus of streptokinase (SK: Ile1-Lys414). It was expressed in a mouse hybridoma cell line and purified by affinity chromatography on a 59D8-antigen column. Chimeric 59D8-SK is a disulfide-linked heterodimer composed of an antibody light chain (Mr 27,000) and a N-glycosylated chimeric heavy chain (M(r) 90,000). The fibrin targeting by 59D8 increased plasma clot lysis by 2-fold, but connecting 59D8 to SK has provided 59D8-SK several unique properties: (i) 59D8-SK activated human Glu-plasminogen with a significant lag period that coincided with limited proteolysis of 59D8-SK similar to that observed for wild-type SK. In a kinetic study, both gave very similar kinetic parameters for the activation of Glu-plasminogen even though 59D8-SK was N-glycosylated in its SK portion; (ii) 59D8-SK was relatively inactive in human plasma, compared to SK, but it became activated in the presence of clots; (iii) 59D8-SK lysed clots slowly but completely whereas SK lysed clots rapidly but incompletely. Even though the mechanism behind these new properties is not fully understood, they are characteristics of a second-generation plasminogen activator.

Design of heterotetrameric coiled coils: evidence for increased stabilization by Glu(-)-Lys(+) ion pair interactions.

Electrostatic interactions between charged amino acids often affect heterospecificity in coiled coils as evidenced by the interaction between the oncoproteins, fos and jun. Such interactions have been successfully exploited in the design of heteromeric coiled coils in a number of laboratories. It has been suggested that heterospecificity in these dimeric coiled-coil systems is driven not by specific electrostatic interactions in the heterodimers but rather by electrostatic repulsion acting to destabilize the homodimer state relative to the heterodimer state. We show that it is possible to design ion pair interactions that directly stabilize the heterotetrameric coiled-coil state. Synthetic peptides were used whose sequences are based on the C-terminal tetramerization domain of Lac repressor, as a model system for four-chain coiled coils (Fairman et al., 1995). These Lac-based peptides, containing either glutamic acid (Lac21E) or lysine (Lac21K) at all b and c heptad positions, only weakly self-associate but, when mixed, afford a highly stable heterotetramer. This study represents the first experimental evidence for the importance of the b and c heptad positions to the stability of coiled coils. Finally, pH dependence and NaCl dependence studies show that heterotetramer stability is driven by ion pair interactions between glutamate and lysine; these interactions contribute about 0.6 kcal/mol of stabilizing free energy for each potential glutamate-lysine pair.

Thrombin receptor activation by thrombin and receptor-derived peptides in platelet and CHRF-288 cell membranes: receptor-stimulated GTPase and evaluation of agonists and partial agonists.

Thrombin receptor activation, by thrombin or SFLLR-containing peptides, stimulates GTPase activity in platelet and CHRF-288 membranes. Polyclonal antibodies to peptides derived from the thrombin receptor (anti-TR52-69 and anti-TR36-49), which block many of thrombin's actions on platelets and endothelial cells, also block thrombin activation of membrane GTPase (as does thrombin active site and anion-binding exosite inhibitors). Most of the receptor-activated GTPase, stimulated by both thrombin and SFLLRNP in platelet membranes, was inhibited by prior treatment with pertussis toxin or N-ethylmaleimide, suggesting that under these conditions much of the thrombin receptor-stimulated GTPase in platelet membranes is a member of the pertussis toxin-sensitive G alpha i family. In platelet membrane preparations, the peptide agonists stimulated approximately twice as much GTPase activity as stimulated by alpha-thrombin. In contrast, the membranes prepared from CHRF-288 cells showed similar maximal SFLLRNP- and alpha-thrombin-stimulated GTPase activity. Stimulation of the platelet membrane GTPase by a variety of different peptide agonists correlated with their ability to stimulate platelet aggregation. Several peptide-based agonists were more potent than the wild-type sequence. The most potent was Ser-(p-fluoro-Phe)-(2-Napthyl-Ala)-Leu-Arg-NH2, which stimulated platelet aggregation (EC50 = 80 nM) and GTPase activity (EC50 = 110 nM). The peptide YFLLRN stimulated GTPase activity but only to approximately 40% of the activity observed with optimal concentrations of other receptor agonists. YFLLRN also limited the stimulation observed with SFLLRNP in a competitive fashion, indicating that YFLLRN is a competitive partial agonist at the thrombin receptor. These studies show that the tethered-ligand receptor mediates the GTPase activation by thrombin in platelet and CHRF-288 cell membranes, and this provides a specific, reliable, and convenient cell-free assay system with which one can evaluate agonists and partial agonists.

A monoclonal antibody against a peptide sequence of fibrinogen gamma chain acts as an inhibitor of factor XIII-mediated crosslinking of human fibrin.

Recurrent hemorrhage has been reported in humans as a result of acquired antibody inhibitors which interfere with the crosslinking of fibrin by factor XIII. One type of these inhibitors (Type III) prevents activated factor XIII from acting on fibrin. We have generated an antifibrin monoclonal antibody, called mAb 4A5, which binds to a peptide sequence at the carboxyl-terminus of human fibrinogen gamma-chains. MAb 4A5 acts like a Type III inhibitor and prevents proper factor XIII-mediated crosslinking. Pre-incubation of fibrinogen or pooled human plasma with mAb 4A5, but not mAb D2 (specific for the carboxyl terminus of fibrin alpha-chains), resulted in clots which are soluble in either 5 M urea or 1% monochloroacetic acid. SDS-PAGE and immunoblotting analysis of these clots confirmed that mAb 4A5 inhibited gamma-chain crosslinking in plasma clots and fibrin clots. Results from a factor XIII activity assay demonstrated that biotinylcadaverine crosslinking into fibrin by factor XIII could be inhibited by mAb 4A5 but not mAb D2, arguing that mAb 4A5 acted by binding the crosslinking site of factor XIII. Studies of the immunoreactivity of these mAbs with 12 different animal species showed that the gamma-chain epitope recognized by mAb 4A5 was more conserved than the alpha-chain epitope recognized by mAb D2. The species fibrinogens, recognized by mAb 4A5 in binding assays, also showed impaired crosslinking when mAb 4A5 was present during the clotting reaction

Characterization of a new four-chain coiled-coil: influence of chain length on stability.

Limited information is available on inherent stabilities of four-chain-coils. We have developed a model system to study this folding motif using synthetic peptides derived from sequences contained in the tetramerization domain of Lac repressor. These peptides are tetrameric as judged by both gel filtration and sedimentation equilibrium and the tetramers are fully helical as determined by CD. The four-chain coiled-coils are well folded as judged by the cooperativity of thermal unfolding and by the extent of dispersion in aliphatic chemical shifts seen in NMR spectra. In addition, we measured the chain length dependence of this four-chain coiled-coil. To this end, we developed a general procedure for nonlinear curve fitting of denaturation data in oligomeric systems. The dissociation constants for bundles that contain alpha-helical chains 21, 28, and 35 amino acids in length are 3.1 x 10(-12), 6.7 x 10(-23), and 1.0 x 10(-38) M3, respectively. This corresponds to tetramer stabilities (in terms of the peptide monomer concentration) of 180 microM, 51 nM, and 280 fM, respectively. Finally, we discuss the rules governing coiled-coil formation in light of the work presented here.

Syk protein-tyrosine kinase is regulated by tyrosine-phosphorylated Ig alpha/Ig beta immunoreceptor tyrosine activation motif binding and autophosphorylation.

Syk is a cytoplasmic protein-tyrosine kinase containing two amino-terminal Src homology 2 domains that is activated following ligation of the B cell antigen receptor. Syk activation in B cells correlates with Syk tyrosine phosphorylation as well as with Syk SH2-mediated association with the tyrosine-phosphorylated Ig alpha and Ig beta B cell antigen receptor subunits. Tyrosine-phosphorylated peptide 20-mers representing Ig alpha and Ig beta immunoreceptor tyrosine activation motifs were synthesized and found to stimulate the specific activity of Syk by as much as 10-fold in vitro. Maximal phosphopeptide-induced Syk activation required both Syk SH2 domains and phosphorylation of both tyrosine residues present in the immunoreceptor tyrosine activation motif. The biochemical mechanism responsible for the phosphopeptide-induced Syk enzyme activation appears to be a function of Syk autophosphorylation. Our observations suggest the association of Syk tandem SH2 domains with the tyrosine-phosphorylated Ig alpha and/or Ig beta immunoreceptor tyrosine activation motifs in B cells stimulates Syk autophosphorylation leading to Syk enzyme activation.

A microtiter assay for factor XIII using fibrinogen and biotinylcadaverine as substrates.

Efforts to develop an improved assay for plasma and tissue transglutaminase have led us to a convenient, sensitive microtiter plate assay for coagulation factor XIII using human fibrinogen as an immobilized substrate. Factor XIII was activated in the presence of calcium, thrombin, and immobilized fibrinogen and then assayed by adding biotinylcadaverine. The reaction was terminated by adding EDTA and the level of incorporated biotin was measured with streptavidin-beta-galactosidase. In this assay, the analytical range for human platelet factor XIII was 0.01-100 ng and 1-100 ng for guinea pig liver transglutaminase. Fibrinogen-coated plates gave more than 100-fold increase in sensitivity compared with N,N-dimethylcasein-coated microtiter plates. The intraassay coefficient of variation was less than 5% (n = 12) and interassay less than 6% (n = 4). The sensitivity of this assay reduced the volumes of plasma samples required and consequently eliminated the need to remove fibrinogen from such test samples. As expected, factor XIII activity could be inhibited by putrescine and antibodies against factor XIII as well as by a monoclonal antibody that bound to the carboxyl terminus of human fibrin gamma-chains. The assay provided a sensitive, simple, and rapid method for measuring factor XIII.

Design and evaluation of a thrombin-activable plasminogen activator.

A new chimeric plasminogen activator with high fibrin affinity was designed to bind fibrin and to initiate clot destruction, following activation by thrombin. The chimeric activator, 59D8-scuPA-T, was made from the Fab fragment of an anti-fibrin antibody (59D8) and a C-terminal portion of a thrombin-activable low molecular weight single-chain urokinase plasminogen activator, scuPA-T, obtained by deletion of Phe-157 and Lys-158 from low molecular weight single-chain urokinase-type plasminogen activator (scuPA) by site-directed mutagenesis. The chimeric molecule had a molecular mass of 91,000, a value consistent with one 59D8 light chain (M(r) = 27,000) and one 59D8 heavy-chain Fd fragment fused to low molecular weight scuPA (M(r) = 64,000). According to its design, 59D8-scuPA-T was activated by thrombin but not by plasmin, whereas the control chimeric molecule, 59D8-scuPA, was activated by plasmin but not by thrombin. When activated by thrombin, 59D8-scuPA-T converted plasminogen to plasmin. In vitro plasma clot lysis assays showed that 59D8-scuPA-T lysed clots performed by thrombin and that heparin and hirudin could prevent clot lysis. When incorporated as part of a thrombin-induced clot, only 59D8-scuPA-T was able to lyse the clot while 59D8-scuPA and high molecular weight scuPA were ineffective. Together these results demonstrate that 59D8-scuPA-T is a thrombin-activable plasminogen activator that offers selective thrombolysis of thrombin-rich clots over more established, aged clots, and may also act as an antithrombotic agent.

Role of histidine 373 in the catalytic activity of coagulation factor XIII.

Factor XIII catalysis proceeds via formation of thioester acyl enzyme intermediate involving an active site cysteine residue at position 314. The contribution of other residues to catalysis has not been established. Earlier studies of the pH dependence of factor XIII activity suggested the existence of a putative active site histidine. We used chemical modification and oligonucleotide directed site-specific mutagenesis to investigate the role of histidines. Photo-oxidation with methylene blue resulted in a complete loss of catalytic activity under conditions that oxidized histidine but did not affect the essential cysteine. Single substitution of each of the 14 histidine residues in the a-subunit of factor XIII by asparagine or alanine led to mutants with catalytic activities generally not significantly different from the wild-type recombinant enzyme. The only exceptions were the H373N and H373A mutants that were poorly expressed, had no detectable rate of [14C]putrescine incorporation into dimethylcasein, and failed to cross-link fibrin gamma-chains. Thus, the a-subunit His-373 may function in the active site of factor XIII, by analogy with papain's mechanism, as a histidinium cation that increases the nucleophilicity of the essential Cys-314. Decreased expression levels of His-373 mutants also indicate that this residue may be critical for enzyme stability.

Characterization of a monoclonal antibody directed against the carboxyl-terminus of human factor XIII. An epitope exposed upon denaturation and conserved across species lines.

By deriving an anti-peptide monoclonal antibody, mAb 7A4, we characterized the relatively unstudied carboxyl-terminal end of the a-chain of human factor XIII, the plasma transglutaminase. MAb 7A4 was directed against the last eight amino acids (Gln-Ile-Gln-Arg-Arg-Pro-Ser-Met) and bound with a dissociation constant of 3.4 x 10(-8)M. In a solid assay format, mAb 7A4 bound equally well to factor XIII obtained from human plasma, platelets or placenta. However, in a solution-phase assay format, the epitope was largely unavailable but could be readily exposed by heat denaturation. Immunoblotting showed that this epitope is conserved among all species of plasma factor XIII tested except rabbit suggesting that the carboxyl-terminus might be an important structural element. Other competitive binding experiments with synthetic peptides as inhibitors pointed toward the final carboxyl-terminal amino acid, Met-731, as an immunochemically important determinant. This was used advantageously to confirm the finding that the carboxyl-terminal Met-731 is largely absent from placental factor XIII (1) as compared to platelet or plasma factor XIII.

Synthetic peptide immunogens for the development of a cardiac myosin light chain-1 specific radioimmunoassay.

To determine if the presence of cardiac light chains in blood could be used to detect acute myocardial infarction, we developed a specific light chain immunoassay. A synthetic peptide sequence specific for human cardiac ventricular myosin light chain 1 (VLC1) was synthesized and designated P348. This peptide coupled to keyhole limpet hemocyanin was used as an immunogen to obtain murine monoclonal antibodies specific for VLC1. Five monoclonal antibodies were obtained. One of these designated Mab-8E3 reacted equally well with both the synthetic peptide and VLC1. Although the 8E3 antibody is specific for VLC1, the use of HPLC purification of skeletal muscle myosin light chain 1 demonstrated that VLC1 is present in human skeletal muscle. The clinical utility of the assay was tested in 18 patients with creatine kinase (CK) and ECG documented acute myocardial infarction. VLC1 was below the limit of detection (< 1 ng/ml) in sera obtained from healthy volunteers and patients without myocardial infarction or chest pain. In contrast VLC1 was elevated in the serum of all 18 patients with acute myocardial infarction. Combining the two test results at the time of admission resulted in 83% of patients having detectable serum levels of one or both markers.

Characterization of adhesion of "resting" and stimulated platelets to fibrinogen and its fragments.

Adhesion of resting and stimulated platelets to immobilized fibrinogen (Fg) was characterized using various forms of Fg, receptor peptide mimics, and antibodies to glycoprotein (GP) IIb/IIIa and Fg. Resting platelets adhered to Fg, but to less than half the extent of the same platelets stimulated with epinephrine/ADP. The adhesion of resting and stimulated platelets to Fg was inhibited by a receptor peptide mimic (G13, a peptide corresponding to residues 300-312 of GPIIb), anti-GPIIb/IIIa antibodies, and a monoclonal antibody (4A5) against the carboxyl terminus of the gamma chain of Fg. The results presented here demonstrate that the alpha chain RGD platelet recognition sites are not required to mediate the adhesion of either stimulated or resting platelets to immobilized Fg. Although stimulated platelets can adhere extensively to monomeric Fg containing one functional gamma chain, resting platelets require bivalent Fg containing two functional gamma chains to mediate irreversible adhesion to Fg.

Selection of monoclonal antibodies that bind and inhibit tissue-type plasminogen activator.

Three monoclonal antibodies raised against tissue-type plasminogen activator (t-PA) were selected for their ability to inhibit solid-phase bound t-PA. Each monoclonal antibody blocked the release of p-nitroaniline from H-D-Ile-Pro-Arg-pNA (S-2288). The first antibody 1D2 was a gamma 2b, kappa with KD = 8 x 10(-9) M, the second antibody 2B9 was a gamma 1, kappa with KD = 2 x 10(-9) M, and the third antibody 5A9 was a gamma 1,kappa with KD = 4 x 10(-10) M. In solution-phase format each antibody blocked the conversion of plasminogen to plasmin as judged by a plasmin assay and also inhibited t-PA-mediated lysis of plasma fibrin clot in plasma. The binding of each 125I-radiolabeled antibody to t-PA was inhibited by any one of the three antibodies, suggesting that they recognized a common epitope on t-PA which was absent on unfolded t-PA. We concluded these antibodies bind near t-PA active site since PPACK treatment lowered binding of two antibodies. We believe solid-phase chromogenic substrate assay may be a useful way to screen for antibodies directed against the active site of proteases.

Activated platelets form protected zones of adhesion on fibrinogen and fibronectin-coated surfaces.

Leukocytes form zones of close apposition when they adhere to ligand-coated surfaces. Because plasma proteins are excluded from these contact zones, we have termed them protected zones of adhesion. To determine whether platelets form similar protected zones of adhesion, gel-filtered platelets stimulated with thrombin or ADP were allowed to adhere to fibrinogen- or fibronectin-coated surfaces. The protein-coated surfaces with platelets attached were stained with either fluorochrome-conjugated goat anti-human fibrinogen or anti-human fibronectin antibodies, or with rhodamine-conjugated polyethylene glycol polymers. Fluorescence microscopy revealed that F(ab')2 anti-fibrinogen (100 kD) did not penetrate into the contact zones between stimulated platelets and the underlying fibrinogen-coated surface, while Fab antifibrinogen (50 kD) and 10 kD polyethylene glycol readily penetrated and stained the substrate beneath the platelets. Thrombin- or ADP-stimulated platelets also formed protected zones of adhesion on fibronectin-coated surfaces. F(ab')2 anti-fibronectin and 10 kD polyethylene glycol were excluded from these adhesion zones, indicating that they are much less permeable than those formed by platelets on fibrinogen-coated surfaces. The permeability properties of protected zones of adhesion formed by stimulated platelets on surfaces coated with both fibrinogen and fibronectin were similar to the zones of adhesion formed on fibronectin alone. mAb 7E3, directed against the alpha IIb beta 3 integrin blocked the formation of protected adhesion zones between thrombin-stimulated platelets and fibrinogen or fibronectin coated surfaces. mAb C13 is directed against the alpha 5 beta 1 integrin on platelets. Stimulated platelets treated with this mAb formed protected zones of adhesion on surfaces coated with fibronectin. These protected zones were impermeable to F(ab')2 antifibronectin but were permeable to 10 kD polyethylene glycol. These results show that activated platelets form protected zones of adhesion and that the size of molecules excluded from these zones depends upon the composition of the matrix proteins to which the platelets adhere. They also show that formation of protected zones of adhesion by platelets requires alpha IIb beta 3 integrins while the permeability properties of these zones of adhesion are regulated by both alpha IIb beta 3 and alpha 5 beta 1 integrins.

The B cell immune response to an idiotype-inducing peptide epitope can be inhibited by immunodominance of a neighboring epitope.

Three synthetic peptides, EEYEYE (peptide 1), EEYEYEEEYEYE (peptide 2), and YEEEEY (peptide 3), were tested for their ability to induce common Id that had been previously characterized for murine antibodies specific to random synthetic polymers of glutamic acid, alanine, and tyrosine ((Glu,Ala,Tyr)n) (cGAT Id). Protein conjugates of either peptide 2 or peptide 3, but not peptide 1, induced cGAT Id. This unique approach directly identified two peptides capable of inducing cGAT Id antibodies. Previous immunization with peptide 1-protein conjugate inhibited the cGAT Id response to peptide 2 conjugated to a different protein carrier. Thus, a neighboring or overlapping epitope can be used to inhibit a cGAT Id-inducing epitope without the participation of carrier-specific Ts and without using anti-Id antibodies. In contrast, previous immunization with peptide 1-protein conjugate did not inhibit cGAT Id induction by peptide 3-protein conjugate or by (Glu,Ala,Tyr)n. This ruled out the participation of Id-specific Ts cells. The effectiveness of inhibition coincided with the avid binding of anti-peptide 1 antibodies to peptide 2, which was > 10 and 100 times stronger than the binding to peptide 3 and (Glu,Ala,Tyr)n, respectively. We hypothesize that the primed peptide 1-specific B cells capture and process peptide 2-protein efficiently and act as APC to Th cells specific to the protein of the challenging Ag, resulting in the selective and dominant activation of peptide 1-specific B cells. Thus, our data suggest that epitope priming can inhibit an Id response to a neighboring epitope by a mechanism of clonal dominance.

Early cross-linked fibrin in human plasma contains alpha-polymers with intact fibrinopeptide A.

alpha-polymer formation, as opposed to gamma-chain dimerization has been considered a relatively late event in factor XIII-induced fibrin stabilization. Recently it has been shown, however, that plasma from healthy individuals and from patients with fibrinaemia contains small amounts of soluble fibrin/fibrinogen oligomers interlinked through dimerized gamma-chains as well as cross-linked alpha-chains. The present work was carried out to see if these early alpha-chain polymers also arise during coagulation of plasma in vitro. Plasma samples from healthy individuals, prepared by immediate centrifugation of blood collected without anticoagulant, were allowed to clot spontaneously for varying periods. The plasma clots were solubilized in SDS-urea-mercaptoethanol and samples were subjected to SDS-PAGE and Western blotting using polyclonal antibodies to human fibrinogen, or monoclonal antibodies specific either for A alpha/alpha-chains, for fibrinopeptide A-containing chains, for the N-terminus of the fibrin beta-chain or for the gamma-chains. Fibrin/fibrinogen oligomers were seen to form long before visible gelation of plasma. These oligomers were cross-linked through gamma-chain dimerization, but also through A alpha- or alpha-chain polymerization. The number and amount of alpha-polymers containing A alpha-chains increased immediately after clot formation, but these disappeared about 20 min later, due to complete removal of fibrinopeptide A (FPA) by thrombin. It is concluded that alpha-polymer formation is a very early event during plasma coagulation in vitro, and that both A alpha- and alpha-chains are involved.

A beta-turn is present in the 392-411 segment of the human fibrinogen gamma-chain. Effects of structural changes in this segment on affinity to antibody 4A5.

The interaction between fibrinogen gamma-peptide 392-411, LTIGEGQQHHLGGAKQAGDV, and monoclonal antibody 4A5, an antibody with a high affinity for both for the peptide and native fibrinogen, is being studied as a model for peptide-antibody interaction. Two-dimensional NMR studies of the free peptide at pH 5.2 indicated the presence of a significant population, about 60%, of type II beta-turn, spanning residues Gln407-Asp410. At pH 2.7, little, if any, turn structure is present. The D-Ala409 analog, which, for steric reasons, would be expected to preserve the beta-turn, and the L-Ala409 analog, which would not be expected to have this conformational feature, were synthesized, and NMR studies confirmed the respective structural predictions. The affinity of the D-Ala analog for antibody 4A5 is even greater than that displayed by native gamma 392-411, while the affinity of the L-Ala analog is less than one-tenth that of the native peptide. Both conformational and steric effects involving residues 407-410 may be important in recognition by antibody 4A5. Since gamma 392-411 includes a platelet receptor binding locus of fibrinogen, and this and related peptides are inhibitors of platelet aggregation, the D-Ala409 and L-Ala409 analogs were tested for platelet binding. Neither of the analogs displays any measurable platelet binding, indicating that the recognition requirements for the platelet receptor differ considerably from those for antibody 4A5.

Immunoelectrophoretic and immunohistochemical characterizations of fibrinogen derivatives in atherosclerotic aortic intimas and vascular prosthesis pseudo-intimas.

Cadaveric aortic intimas with uncomplicated atherosclerosis were examined to determine the distribution and polypeptide chain composition of fibrinogen-related protein. Immunohistochemical staining showed deposits rich in fibrinopeptides A and B. The deposits were usually disseminated throughout intimas of moderate thickness < 0.7 mm, but were distributed focally in elongate patches bounded both lumenally and medially by deposit-free tissue in thick atheromas. Saline extracts generally showed undegraded monomers and dimers by electrophoresis. The residual protein contained A alpha and gamma-chains that were cross-linked predominantly (>80%) into unresolved high M(r) (>200 kd) derivatives, whereas B beta-chains were left non-cross-linked, as occurs in late stages of cross-linking by transglutaminases. The resolved components had electrophoretic mobilities corresponding to characteristic products of both factor XIIIa and tissue-transglutaminase. A greater incorporation of alpha- rather than gamma-chains into cross-linked products implicated tissue-transglutaminase as contributing heavily. By contrast, vascular graft pseudo-intimas and a cadaveric clot were rich in degraded fibrin devoid of fibrinopeptide A, and cross-linked in patterns typical of XIIIa with gamma 2 dimers constituting the principal product. The findings indicate that the fibrinogen in the aortic intima is comparatively well protected from thrombin and plasmin, and that much of it is deposited through direct cross-linking by tissue-transglutaminase without being converted to fibrin.