Contact with the N termini in the central E domain enhances the reactivities of the distal D domains of fibrin to factor XIIIa.

The reaction of Factor XIII with fibrin is the last enzyme-catalyzed step on the coagulation cascade, leading to the formation of a normal blood clot. The finding that fibrin is preferred by the cross-linking enzyme about 10-fold over the circulating fibrinogen suggests the operation of a unique substrate-level control for the orderly functioning of the physiological process in the forward direction. An important task is to elucidate the molecular mechanism for the transmission of the signal generated by the thrombin-catalyzed cleavage in the central E domain of fibrin to the distant Factor XIII-reactive glutamine residues. By focusing on the substrate sites present in chain remnants of D type domains of fibrinogen and by employing the approach of fragment complementation with the regulatory E domain, which represents the thrombin-modified portion of fibrin, we have now succeeded in reconstructing in solution the phenomenon of kinetic enhancement for the reaction with Factor XIII. Two D type preparations (truncated fibrinogen, 250 kDa and D‘, 105 kDa) were obtained by digestion of human fibrinogen with endo Lys-C. Neither product could be cross-linked by Factor XIII, but as shown by the incorporation of dansylcadaverine, both were acceptor substrates for the enzyme. The plasmin-derived D (105-kDa) product, however, could be cross-linked into DD dimers. In all cases, the admixture of E fragments exerted a remarkable boosting effect on the reactions with Factor XIII. Even with native fibrinogen as substrate, cross-linking of chains was enhanced in the presence of E. Nondenaturing electrophoresis was used to demonstrate the complex forming potential of E fragments with fibrinogen, truncated fibrinogen, D‘, or D. The GPRP tetrapeptide mimic of the GPRV N-terminal sequence of the α chains in the E fragments, abolished both complex formation and the kinetic boosting effect of E on the reactions of substrates with Factor XIII. Thus, the N-terminal α chain sequences seem to act as organizing templates for spatially orienting the D domains, probably during the protofibrillar assembly of the fibrin units, for favorable reaction with Factor XIII.

The reaction of Factor XIII a with fibrin is the last enzyme-catalyzed step on the coagulation cascade, leading to the formation of a normal blood clot. The finding that fibrin is preferred by the cross-linking enzyme about 10-fold over the circulating fibrinogen suggests the operation of a unique substrate-level control for the orderly functioning of the physiological process in the forward direction. An important task is to elucidate the molecular mechanism for the transmission of the signal generated by the thrombin-catalyzed cleavage in the central E domain of fibrin to the distant Factor XIII areactive glutamine residues. By focusing on the substrate sites present in ␥ chain remnants of D type domains of fibrinogen and by employing the approach of fragment complementation with the regulatory E domain, which represents the thrombin-modified portion of fibrin, we have now succeeded in reconstructing in solution the phenomenon of kinetic enhancement for the reaction with Factor XIII a .
Two D type preparations (truncated fibrinogen, ϳ250 kDa and D, ϳ105 kDa) were obtained by digestion of human fibrinogen with endo Lys-C. Neither product could be cross-linked by Factor XIII a , but as shown by the incorporation of dansylcadaverine, both were acceptor substrates for the enzyme. The plasmin-derived D (ϳ105-kDa) product, however, could be cross-linked into DD dimers. In all cases, the admixture of E fragments exerted a remarkable boosting effect on the reactions with Factor XIII a . Even with native fibrinogen as substrate, cross-linking of ␥ chains was enhanced in the presence of E. Nondenaturing electrophoresis was used to demonstrate the complex forming potential of E fragments with fibrinogen, truncated fibrinogen, D, or D. The GPRP tetrapeptide mimic of the GPRV N-terminal sequence of the ␣ chains in the E fragments, abolished both complex formation and the kinetic boosting effect of E on the reactions of substrates with Factor XIII a . Thus, the N-terminal ␣ chain sequences seem to act as organizing templates for spatially orienting the D domains, probably during the protofibrillar assembly of the fibrin units, for favorable reaction with Factor XIII a .
Stabilization of the fibrin network during the last phase of blood coagulation by the Factor XIII system represents a tightly regulated sequence of events. Though clotting times may be normal, failures in regard to the functioning of any aspect of this system could give rise to severe and frequently fatal bleeding (1,2). The introduction of N ⑀ (␥-glutamyl)lysine cross-links by activated Factor XIII (or Factor XIII a ) strengthens the clot (3)(4)(5) and renders it more resistant to lytic enzymes (6,7). Until now, attention was paid mostly to the biochemical regulation of the conversion of circulating Factor XIII (or fibrin stabilizing factor A 2 B 2 ) to Factor XIII a (or A 2 *). Activation of this zymogen occurs in two distinct stages, catalyzed by thrombin and promoted by Ca 2ϩ , respectively: (8,9).
Fibrin, which is the physiological substrate of the crosslinking enzyme, acts as a "feed forward" modulator in the conversion (10 -15), ensuring that the transamidating enzyme (A 2 *) is produced within the physiologically required time frame for efficient clot stabilization. Factor XIII a reacts first with the cross-linking sites in the ␥ chains of fibrin and then with those in the ␣ chains (16 -18).
The known preferential reactivity of Factor XIII a for fibrin over fibrinogen (17,19) is considered to be yet another regulatory feature for the normal functioning of the clot-stabilizing system. The thrombin-induced alteration in the central N-terminal domains of fibrinogen (or, more precisely, the mere cleavage of fibrinopeptide A from the A␣ chains (17)) must generate the signal for activating the distant, Factor XIII a -reactive sites in the protein. The purpose of the present work is to shed light on this unique substrate level modulation in blood coagulation. By focusing on the ␥ chain cross-linking sites of the substrate, we succeeded in reconstructing in solution the regulatory mechanism by complementing fragments of the Factor XIII areactive distal D regions of fibrinogen with the central thrombin-modified E domains of fibrin.

MATERIALS AND METHODS
Human Factor XIII was purified from outdated blood bank plasma (10,20,21) and was stored for up to 3 months at 4°C at a concentration of 1 mg of protein/ml in 50 mM Tris-HCl, 1 (22). A truncated form of fibrinogen was prepared by digestion with endoproteinase Lys-C (Promega, Madison, WI) by a procedure similar to that described for bovine fibrinogen (23). Human fibrinogen (50 mg) was dissolved in 5 ml of TBS and dialyzed against 2 liters of the same overnight at 4°C. Then 25 l of 1 M CaCl 2 was added to a concentration of 5 mM, and the mixture was warmed to 37°C 30 min prior to adding 5 l of 20 units/ml endoproteinase Lys-C (final concentration of 0.02 units/ml). Following incubation at 37°C for 2.5 h, 50 l of 100 mM N ␣ -p-tosyl-L-lysine chloromethyl ketone (Sigma) in TBS was added, to a concentration of 1 mM. The digest was dialyzed against 2 liters of TBS (overnight, 4°C), and the truncated fibrinogen was purified by gel filtration on an Ultrogel AcA 44 column (LKB, Reactifs IBF, Pharmindustrie, France; 2.6 x 87 cm, equilibrated with 0.1 M NH 4 HCO 3 , pH 8.0, and eluted with a flow rate of 30 ml/h at room temperature). Fractions of 5 ml were collected and analyzed by SDS-PAGE (24); those eluting with the void volume were pooled, lyophilized, dissolved in TBS, and dialyzed at 4°C against the latter overnight. The protein concentration of the product was determined by the bicinchoninic acid assay (Pierce; Ref. 25) using human fibrinogen as reference, and the material (8.0 mg/ml) was stored at Ϫ20°C. The preparation was free of the intact fibrinogen starting material and comprised one major protein band of about 250 kDa in SDS-PAGE under nonreducing conditions. In reducing SDS-PAGE (40 mM dithiothreitol (Sigma) in the sample buffer, 100°C, 3 min) three components were found with molecular masses of 48, 42, and 28 kDa, respectively. The 48-kDa band was actually a doublet. When examined by Western blotting (26)  The distal D fragment of fibrinogen was obtained by activation of endogenous plasminogen with urokinase (109,127 units/mg, Calbiochem, La Jolla, CA) according to some modifications of a previously published procedure (30). To 5 ml of human fibrinogen solution (50 mg of protein in TBS, exhaustively dialyzed at 4°C), 25 l of 1 M CaCl 2 was added. After warming to 37°C (30 min), 5 l of 5,000 units/ml human urokinase was admixed, followed by incubation at 37°C for 80 min. To terminate plasmic digestion, Trasylol (100 l of 10,000 kallikrein-inactivating units/ml) was added, and the mixture was dialyzed against 500 ml of TBS at 4°C for 2 days. Purification of the D fragment was accomplished by gel filtration on an Ultrogel AcA 44 column as described above for truncated fibrinogen with inclusion of 5 kallikreininactivating units/ml of Trasylol in the eluting buffer. Fractions of 5 ml were collected. When analyzed by nonreducing SDS-PAGE (24), the first emerging protein peak was found to comprise intact fibrinogen, X and Y fragments (31)(32)(33)(34), whereas the early portion of the second major peak showed a mixture of D and E fragments (32)(33)(34)(35). The last three fractions of this peak, however, contained D fragments alone. These fractions were pooled, lyophilized, and dialyzed as described for truncated fibrinogen. The preparation (11.8 mg/ml, bicinchoninic acid assay; Pierce (25)), comprising approximately equal amounts of D 1A and D 1 (36) as judged by nonreducing SDS-PAGE, was stored at Ϫ20°C.
The central domain E fragment was purified from cross-linked fibrin according to some modifications of a previously published procedure (37)(38)(39). To 10 ml of human fibrinogen (100 mg of protein in TBS, exhaustively dialyzed at 4°C), 50 l of 1 M CaCl 2 was added. After warming to 37°C (30 min), 5 l of 1 mg/ml of human Factor XIII and 10 l of 1,000 NIH units/ml of human ␣-thrombin was admixed, followed by incubation at 37°C overnight. Human plasminogen (20.7 casein units/mg, Chromogenix AB (Molndal, Sweden)) was activated by mixing 4 l of 1 mg/ml solution in TBS with 0.5 l of 5,000 units/ml of human urokinase followed by incubation for 1 h at 37°C. The cross-linked clot was dispersed, 2.5 l of activated plasminogen solution was added, and the mixture was incubated overnight at 37°C on a shaker. Trasylol (25 l of 10,000 kallikrein-inactivating units/ml solution) was added, and the digest was dialyzed against 2 liters of TBS overnight at 4°C. Purification of the DD⅐E complex was accomplished by gel filtration on an Ultrogel AcA 44 column as described above for truncated fibrinogen. Fractions of 5 ml were collected. When analyzed by nonreducing SDS-PAGE (24), the first emerging protein peak was found to comprise the DD⅐E complex (37). Pooled and lyophilized fractions (57 mg by weight) were dissolved in 5.7 ml of 150 mM sodium citrate, 3 M urea at pH 5.5 and incubated at 37°C for 5 h. Purification of the dissociated E fragments was accomplished by gel filtration on an Ultrogel AcA 44 column as described above for truncated fibrinogen; however, the column was equilibrated with 50 mM sodium citrate, 3 M urea, pH 5.5. Fractions emerging from the column in the second peak were pooled, dialyzed against 2 ϫ 2 liters of 0.1 M NH 4 HCO 3 , pH 8.0, overnight at 4°C, lyophilized, and dialyzed against TBS as described for truncated fibrinogen. The preparation (15.8 mg/ml, bicinchoninic acid assay; Pierce (25)) was comprised of 39% E 1 , 48% E 2 , and 13% E 3 (40), as determined by nonreducing SDS-PAGE (24). It was stored at Ϫ20°C.
The distal D-like fragment (DЈ) of fibrinogen was also obtained by digestion with endoproteinase Lys-C as described for the truncated fibrinogen, with the exception that the digestion time was increased to 4 h at 37°C, and the fractions emerging from the column in the second peak were pooled and lyophilized. The preparation comprised two protein bands of 110 and 105 kDa in SDS-PAGE under nonreducing conditions. In reducing SDS-PAGE, four components were identified with molecular masses of 45, 42, 16, and 14 kDa, respectively. The purified material (8.6 mg/ml, bicinchoninic acid assay; Pierce (25)) was stored at Ϫ20°C.
Activation of Factor XIII to XIII a by thrombin and Ca 2ϩ was carried out at room temperature for 30 min using 50 NIH units of thrombin/mg of zymogen. Thus, a typical mixture sufficient for multiple experiments (36 l) was comprised of 4 l of Factor XIII (1 mg/ml stock), 4 l of thrombin (50 NIH units/ml stock), 26 l of TBS, and 2 l of CaCl 2 (100 mM stock) in TBS. At the end of the incubation period, in most experiments, thrombin activity was quenched by addition of 4 l of hirudin (Sigma; 200 units/ml stock). Otherwise, 4 l of thrombin (same stock as above) was added instead of hirudin.

Probing the Factor XIII a Reactive Sites of Truncated Fibrinogen and of the D and DЈ Fragments by Incorporation of Dansylcadaverine-
These experiments were carried out by a modification of a previously published procedure (42)(43)(44). Enzyme-catalyzed incorporation of the amine was terminated by adding 100 l of 20% (w/v) trichloroacetic acid to 20 l of experimental mixtures, followed by the addition of 80 l of fibrinogen (5 mg/ml stock) as a carrier. Protein precipitation was allowed to proceed at room temperature for 10 min, and then more trichloroacetic acid (0.8 ml, 10%) was added and mixed in with Vortex. The precipitate was separated by centrifugation (15,600 ϫ g for 3 min; Eppendorf Microfuge, Brinkmann Instruments, Inc., Westbury, NY). Free dansylcadaverine was removed by repeated (5 ϫ 1-ml) extraction with ethanol:ether (1:1, v/v), and the precipitate was partially dried in air for 10 min. It was then digested overnight at 37°C with trypsin (100 l of 0.1 mg/ml in 0.1 M NH 4 HCO 3 , pH 8.0; Worthington). A solution of SDS (0.5%) and urea (8 M) in TBS was added (1.9 ml), and fluorescence due to the protein-bound dansyl label was measured ( ex ϭ 350 nm, em ϭ 530 nm) in an SLM 8000C spectrofluorometer (SLM Aminco, Urbana, IL) and was calculated against a reference solution (1 M) of dansylcadaverine.
Electrophoresis-Both denaturing and nondenaturing conditions were employed for the analysis of protein profiles. SDS-PAGE (24) was performed with and without reduction of samples by dithiothreitol in a Mini-PROTEAN II system (Bio-Rad). Protein labeling by dansylcadaverine was documented by photographing the gels under UV illumination (366 nm; UVL-56 lamp, UVP, Inc., San Gabriel, CA) prior to staining with 0.25% Coomassie Brilliant Blue R-250 (Sigma) in 25% methanol, 10% acetic acid for 30 min followed by destaining with 5% methanol, 10% acetic acid. Gel scans were obtained in an UltroScan XL (LKB, Bromma, Sweden) apparatus. Nondenaturing electrophoresis was carried out either in the PhastSystem (8 -25% and 4 -15% gradient gels; Pharmacia Biotech Inc.) or in gels with a 4 -20% linear gradient according to Laemmli (24) (0.375 M Tris-HCl, pH 8.8; Bio-Rad), with 25 mM Tris, 192 mM glycine, pH 8.3, as the running buffer. Staining and destaining, as well as scanning of the gels, was carried out as described above.

RESULTS AND DISCUSSION
Reaction of fibrin occurs much faster with Factor XIII a than reaction of the parent fibrinogen molecule (17,19). Depending on individual preparations and clotting conditions, a 4 -15-fold increase in rate is observed for fibrin for the enzyme-catalyzed incorporation of small primary amine probes, such as glycine ethyl ester, hydroxylamine, hydrazine, or dansylcadaverine. Placed in the framework of other physiologically important regulatory processes, this rate differential (matching, for example, the difference in clotting times between a normal and a hemophilic plasma) is of obvious significance. The difference in rates for incorporation of about 2 equivalents of amines reflects mainly on the reactivities of ␥ chain acceptor sites (16), located in the distal D domains of the protein (45). The ␣ chain acceptors, themselves capable of reacting with 4 -6 equivalents of amines become engaged later (16,17,46).
The kinetic advantage for fibrin over fibrinogen can be seen also when, instead of thrombin, a snake venom enzyme, Arvin, is employed for clotting (17). Since Arvin cleaves only fibrinopeptide A from the N termini of the A␣ chains, this limited proteolytic alteration of fibrinogen alone seems to be responsible for triggering the enhanced reactivity of amine acceptor sites for Factor XIII a . However, the relevant N-terminal region of the protein, located in its central E domain, is far removed from the Factor XIII a -reactive glutamine residues (45)(46)(47). This poses the intriguing question as to what the actual mechanism of the long range communication between the two distant loci might be.
In the present work, we focused exclusively on the reactivities of the ␥ chain acceptor sites in the D domains of fibrin-(ogen). Reactivity to Factor XIII a was probed by the incorporation of dansylcadaverine; reaction products were measured quantitatively (42)(43)(44) and were also documented on the SDS-PAGE profiles of the proteins photographically under UV illumination (17). Three partially degraded forms of human fibrinogen were used as substrates. Digestion of the parent protein with endo Lys-C yielded a ϳ250-kDa species, referred to as truncated fibrinogen, and also a ϳ105-kDa fragment, called DЈ, while digestion of fibrinogen with plasmin produced the known D fragments. The latter comprise the A␣- Experiments with Endo Lys-C-truncated Fibrinogen and with Native Fibrinogen-The endo Lys-C digestion product, having lost the amine incorporation sites from the C-terminal two-thirds portion of its A␣ chains, retained only the ␥ chain sites for reaction with Factor XIII a . However, because of a cleavage at Lys-␥406, the truncated fibrinogen cannot form protein-to-protein cross-bridges upon treatment with Factor XIII a . It serves as a pure acceptor substrate for the enzyme, capable of incorporating amines only into the ␥ chain glutamine sites. Moreover, the N termini of the A␣ chains of the truncated fibrinogen are still intact and, thus, this region is available for modification by thrombin. A prime question was whether this truncated product would still be susceptible to the thrombin-dependent substrate level regulation toward Factor XIII a , observed previously for intact fibrinogen (17,19). The results presented in Fig. 1 clearly answer the question in the affirmative. Although the endo Lys-C truncated fibrinogen has a mass about 30% smaller than the parent protein, it still contains the structural framework necessary for the transmission of information from the thrombin-catalyzed cleavage sites at N termini of the A␣ chains to the distal Factor XIII a -reactive sites in the ␥ chains. The rate enhancement for the thrombinmodified truncated fibrinogen substrate (ϳ5-fold) was well within the range observed for native fibrin.
The next issue was to test whether the Factor XIII a reactivity of sites could be enhanced without actually modifying the truncated fibrinogen substrate, itself, by thrombin. In these experiments, a fixed concentration (4 M) of truncated fibrinogen was mixed with varying concentrations (0 -16 M) of the preparation of E fragments, representing the thrombin-modified central domains of fibrin. As seen in Fig. 2, admixture of E produced a dramatic enhancement in the rate of dansylcadaverine incorporation into the Factor XIII a -reactive sites of truncated fibrinogen. This finding, coupled with the observation that the aggregation-inhibitory (52) tetrapeptide GPRP (5 mM; Oz Chemical Co., Israel), abolished the effect, may be taken to indicate that a direct contact with the E domain was required for activating the substrate for Factor XIII a . Maximal rate enhancement was obtained at an approximate 1:1 molar ratio of truncated fibrinogen to E fragment added. The results predicted that truncated fibrinogen would form a complex with the E fragments. Indeed, the existence of such a complex in solution could be readily demonstrated by nondenaturing electrophoresis (Fig. 3).
If E fragments had a similar ability to bind to native fibrinogen, reactivity of the latter substrate for Factor XIII a might also be up-regulated. While some researchers could not find evidence for complex formation between native fibrinogen and E fragments (53), others were able to document the slow pro- duction of complexes of fibrinogen with E and also with N-DSK fragments (a near equivalent of E) growing into large aggregates (54 -59). In reexamining this issue by nondenaturing electrophoresis, using either the Pharmacia PhastSystem gels (Fig. 4) or the Bio-Rad gradient gel system (data not shown), native fibrinogen could definitely be shown to bind E fragments in solution. Moreover, in tune with the observations on truncated fibrinogen (Fig. 2), the rate of the Factor XIII a -catalyzed reaction with native fibrinogen could also be significantly increased by admixture of E fragments without the necessity of including thrombin in the medium (Fig. 5). Again, maximal enhancement of ␥ chain cross-linking for the native fibrinogen was obtained by supplementation with equimolar E, and the GPRP tetrapeptide abolished the augmenting effect of E fragments.
Complementation with E Fragments in Solution Enhances the Reactivities of the D Fragments of Fibrinogen to Factor XIII a -Further insight into the activation of Factor XIII a -reactive ␥ chain sites was sought by studying the behavior of this purified domain of the substrate, obtained by digestion of fibrinogen with either endo Lys-C (yielding DЈ) or plasmin (yielding D). tested for reactivities to Factor XIII a . The experiments presented in Figs. 6 and 7 illustrate our findings. Fig. 6 shows the enhancement obtained for the incorporation of dansylcadaverine into the DЈ fragment upon complementation with E. Maximal rate enhancement was seen at a mixing ratio of about 2 mol of DЈ/mol of E (i.e. 8:4 M). By holding this mixing ratio constant for increasing concentrations of the complex and measuring the initial velocity of amine incorporation (data not shown), an approximately 5-7-fold enhancement in the k cat /K m catalytic efficiency index was calculated for the reaction of the [2DЈ⅐E] complex with Factor XIII a in comparison to that of free DЈ.
Employing the plasmin-derived D fragment of fibrinogen as substrate, the effect of complementation by E could be examined for the formation of ␥-␥ chain cross-linked D-dimers under the influence of Factor XIII a . This reaction, too, was greatly promoted by the admixture of E fragments (Fig. 7). As with the reaction involving DЈ (Fig. 6), the GPRP tetrapeptide abolished the positive modulatory influence of E, indicating that a direct contact with E was necessary to activate the Factor XIII areactive sites present in both DЈ and D fragments.
A review of the literature suggests that, in previous efforts, considerable difficulties have been encountered for directly demonstrating complex formation between monomeric D fragments (from the plasmic digest of fibrinogen) and E in solution. However, in solid phase assays, D could be shown to bind E. Moreover, in mixtures of D with E, products corresponding to 1D⅐1E and 2D⅐1E could be generated through the action of a bifunctional chemical cross-linking agent (30). As presented in Fig. 8, with nondenaturing electrophoresis using the Bio-Rad gradient gel system, we found definite evidence for the existence of complexes of DЈ⅐E and D⅐E in solution. Similar findings were obtained with the Pharmacia Phast System (data not shown).
The specific interactions of substrates with the E fragments must be critical for functionally up-regulating their Factor XIII a -reactive sites. It is important to point out that the addition of E fragments had no influence whatever on the Factor XIII a -catalyzed incorporation of dansylcadaverine into Ndimethylated casein as a test substrate (data not shown).
Formation of the copulation plug in rodents represents another example of a transglutaminase-driven biological clotting phenomenon in which a specific substrate level regulation was found to play a major role. The postejaculatory clotting of proteins discharged from the seminal vesicles is brought about by the Ca 2ϩ -dependent enzyme secreted simultaneously from the anterior lobe of the prostate. Acid glycoproteins derived from the bulbo-urethral (Cowper's) gland greatly accelerate formation of the coagulum (60 -62). In the reconstruction of this clotting system with purified components, macromolecular polyanions could substitute for the Cowper's gland secretion. The polyanions seem to exert a positive modulatory effect by interacting specifically with the seminal vesicle secretion proteins so as to enhance their proclivity for acting as transglutaminase acceptor substrates (63). The situation is quite analogous to the effect exerted by the E fragments for specifically promoting the reactions of the fibrinogen-derived substrates for Factor XIII a .
In tracing the path for the transmission of the up-regulating signal from the thrombin-modified central E domain of fibrin to the Factor XIII a -reactive distal glutamine residues, the following must be borne in mind. The GPRP tetrapeptide, a mimic for the GPRV N-terminal sequence of the ␣ chains of fibrin generated by the cleavage of fibrinopeptide A (52), in our experiments abolished both the complex-forming ability and the ki- netic boosting effect of E on the reactions of D type fragments with Factor XIII a . Thus, the GPRV ligand in E may be construed to be the initiating signal for the observed phenomenon. This N-terminal ␣ chain sequence in E is suggested to act as an organizing template to spatially orient the bound D domains for favorable reaction with Factor XIII a . The E domain, in a tether-like fashion, would promote the noncovalent dimerization of D domains. Such interaction of E with the D domains is known to take place (64) during the protofibrillar assembly of fibrin units into half-staggered arrays (with a periodicity of 46:2 ϭ 23 nm; (64)). The combining site for the GPRV ligand is likely to be in the vicinity of D domain residues involved in binding the competing GPRP tetrapeptide, i.e. near Tyr-363 in the ␥ chains (65)(66). How the effect of the binding of E to D at this site is further transmitted down the line within the D domain to boost the reactivity of Gln-398 in the ␥ chains still remains to be elucidated. It is clear, however, that the E to D contact carries a significant functional consequence for the entire clotting process by boosting the catalytic efficiency for the Factor XIII a -reactive glutamine residues by about an order of magnitude.