Thrombin-activable Fibrinolysis Inhibitor Attenuates (DD)E-mediated Stimulation of Plasminogen Activation by Reducing the Affinity of (DD)E for Tissue Plasminogen Activator

A complex of d-dimer noncovalently associated with fragment E ((DD)E), a degradation product of cross-linked fibrin that binds tissue plasminogen activator (t-PA) and plasminogen (Pg) with affinities similar to those of fibrin, compromises the fibrin specificity of t-PA by stimulating systemic Pg activation. In this study, we examined the effect of thrombin-activable fibrinolysis inhibitor (TAFI), a latent carboxypeptidase B (CPB)-like enzyme, on the stimulatory activity of (DD)E. Incubation of (DD)E with activated TAFI (TAFIa) or CPB (a) produces a 96% reduction in the capacity of (DD)E to stimulate t-PA-mediated activation of Glu- or Lys-Pg by reducing k cat and increasingK m for the reaction; (b) induces the release of 8 mol of lysine/mol of (DD)E, although most of the stimulatory activity is lost after release of only 4 mol of lysine/mol (DD)E; and (c) reduces the affinity of (DD)E for Glu-Pg, Lys-Pg, and t-PA by 2-, 4-, and 160-fold, respectively. Because TAFIa- or CPB-exposed (DD)E produces little stimulation of Glu-Pg activation by t-PA, (DD)E is not degraded into fragment E and d-dimer, the latter of which has been reported to impair fibrin polymerization. These data suggest a novel role for TAFIa. By attenuating systemic Pg activation by (DD)E, TAFIa renders t-PA more fibrin-specific.

sociated with partial degradation of fibrin by plasmin.
Given its mechanism of action, we speculated that TAFIa or CPB would compromise the cofactor activity of (DD)E by abrogating Pg and/or t-PA binding. To explore this possibility, the ability of (DD)E to potentiate Pg activation by t-PA was examined before and after exposure of (DD)E to TAFIa or CPB. Herein we demonstrate that the stimulatory activity of (DD)E is nearly abolished upon exposure to TAFIa or CPB. Because plasmin generation in the presence of TAFIa-or CPB-exposed (DD)E is less than that with native (DD)E, TAFIa or CPB attenuates degradation of (DD)E into its constitutive fragments, E and DD, the latter of which can impair fibrin polymerization. To explore the mechanism responsible for this phenomenon, light scattering spectroscopy was used to compare the affinities of TAFIa-or CPB-exposed (DD)E for Glu-Pg, Lys-Pg, or t-PA with those of native (DD)E. TAFIa or CPB exposure reduces the affinity of (DD)E for Glu-Pg and Lys-Pg 2and 4-fold, respectively. In contrast, the affinity of (DD)E for t-PA is reduced 160-fold. These data raise the possibility that by attenuating the systemic plasmin generation induced by (DD)E, TAFIa renders t-PA more fibrin-specific.

Materials
Plasminogen Activator-Wild-type recombinant t-PA, kindly provided by Dr. B. Keyt (Genentech Inc., S. San Francisco, CA) was 93% single chain when analyzed by SDS-polyacrylamide gel electrophoresis (PAGE) on 4 -15% gradient gels (Ready-Gel; Bio-Rad) as determined using an ImageMaster Video Documentation System (Amersham Pharmacia Biotech). Active site blocked t-PA was prepared by incubating the activator with a 5-fold molar excess of D-phenyl-prolyl-arginine chloromethyl ketone (PPACK; Calbiochem) as described previously (5,20). The molecular weight and extinction coefficient used for t-PA were 65,000 and ⑀ 1% 280 ϭ 20.0, respectively (4). Plasminogen-Native Glu-Pg was isolated from fresh frozen plasma by lysine-Sepharose affinity chromatography as described previously (5,21). Isolated Glu-Pg was free of Lys-Pg based on urea/acetic acid PAGE analysis (22) and contained no plasmin as assessed using the plasmin-directed substrate D-valyl-leucyl-lysine p-nitroanilide dihydrochloride (S-2251; Chromogenix, Mississauga, Canada). Glu-Pg concentrations were calculated by measuring absorbances at 280 and 320 nm to distinguish tryptophan and phenylananine absorbance from light scattering, respectively (23), using a molecular weight of 90,000 and ⑀ 1% 280 ϭ 16.1 (21). Lys-Pg was purchased from Enzyme Research Laboratories Inc. (South Bend, IN). Lys-Pg was free of Glu-Pg and contained no plasmin, determined as described above.
Fibrinogen-Human Pg-depleted Fg was purchased from Enzyme Research Laboratories and rendered fibronectin-free by batch adsorption with gelatin-agarose (Sigma) for 30 min at room temperature followed by centrifugation at 3000 ϫ g for 10 min (11). Fg concentration was calculated by measuring absorbances at 280 and 320 nm and using a molecular weight of 340,000 and ⑀ 1% 280 ϭ 16.0 (24). (DD)E-The fibrin degradation product, (DD)E, was prepared by plasmin-mediated lysis of cross-linked fibrin clots as described previously (11). Briefly, 100 mg of Fg was clotted with 64 nM human ␣-thrombin (Enzyme Research Laboratories) and 10 mM CaCl 2 in the presence of 93 nM activated recombinant factor XIII (a generous gift from Dr. P. Bishop, Zymogenetics Inc., Seattle, WA), 0.4 M Glu-Pg, and 4.3 nM t-PA. Under these conditions, a clot formed within 30 s. After a 20-min incubation, the reaction was terminated by the addition of 5 M D-valylphenyl-lysine chloromethyl ketone (VFKCK; Calbiochem) and 1 M PPACK to inhibit plasmin and t-PA, respectively. Remaining aggregates were removed by centrifugation at 12,000 ϫ g for 5 min, and fibrin degradation products were separated using a Biosep-Sec-S3000 size exclusion column (Phenomenex, Torrence, CA) fitted to a liquid chromatography system (System Gold; Beckman Instruments Inc., Palo Alto, CA). (DD)E-containing fractions were identified based on their apparent molecular weight and by immunoblot analysis using antibodies against DD and fragment E (12). When subjected to PAGE under nondenaturing conditions, (DD)E migrated as a single band. As expected, upon the addition of 10 mM H-Gly-Pro-Arg-Pro-OH (GPRP, Novabiochem, San Diego, CA) prior to native PAGE, or when subjected to SDS-PAGE, two lower molecular weight bands appeared, corresponding to DD and fragment E, respectively (8,25). (DD)E concentrations were calculated by measuring absorbances at 280 and 320 nm and using a molecular weight of 250,000 and ⑀ 1% 280 ϭ 16.0 (8). Carboxypeptidase B and TAFIa-CPB, potato tuber-derived CPB inhibitor (CPI), and the CPB-directed synthetic substrate hippuryl-Larginine were purchased from Sigma. CPB activity was assessed by incubating 20 nM CPB with 0.4 mM hippuryl-L-arginine dissolved in 0.02 M Tris-HCl, 0.15 M NaCl, 0.01% Tween 20, pH 7.4 (TBS) in a quartz cuvette. Increases in absorbance at 254 nm were monitored for 20 min at 22°C using a DU 7400 Spectrophotometer from Beckman (Mississauga, Canada). Under these conditions, CPB has a specific activity of 41 units/mg, where one unit hydrolyzes 1 mol of hippuryl-L-arginine/ min. When the experiment was repeated in the presence of 1 M CPI, the lowest concentration used to inhibit CPB prior to Pg activation assays, no increase in A 254 was observed, indicating complete CPB inhibition. TAFI was isolated from fresh frozen human plasma by Pg-Sepharose affinity chromatography and activated with thrombin and soluble thrombomodulin as described elsewhere (15,26). The specific activity of the resultant TAFIa against hippuryl-L-arginine was similar to that of CPB. Because TAFIa activity is unstable at room temperature (27,28), TAFIa was used immediately or kept on ice until used. Like CPB, the activity of 20 nM TAFIa was completely inhibited by 1 M CPI. CPB was used for the majority of experiments because its activity is more stable than that of TAFIa and it does not require preactivation. To demonstrate that TAFIa has effects similar to CPB, however, confirmatory experiments were done using TAFIa.

Methods
Effect of TAFIa or CPB on the Rate of (DD)E-stimulated Glu-and Lys-Pg Activation by t-PA-To examine the effect of TAFIa or CPB on the ability of (DD)E to stimulate Pg activation by t-PA, 1.0 ml of a 8 M (DD)E solution was incubated with 20 nM TAFIa or CPB for 40 min at 22°C. At intervals, 20-l aliquots were removed, and 2 l of 40 M CPI was added to inhibit the TAFIa or CPB. Complete TAFIa or CPB inhibition was achieved because no residual activity was detected using the CPB-directed synthetic substrate hippuryl-L-arginine. The ability of TAFIa-or CPB-treated (DD)E to stimulate Pg activation was then compared with that of the starting material by adding 20 l of 2 mM S-2251 and 0.5 nM t-PA to wells of a 96-well microtiter plate containing 0.5 M Glu-Pg or 0.12 M Lys-Pg and 0.5 M (DD)E, dissolved in 80 l of TBS. Plasmin generation was assessed by measuring absorbance at 405 nm at 30-s intervals for 60 min using a Spectramax microplate spectrophotometer (Molecular Devices, Sunnyvale, CA) thermostated at 37°C. Initial rates of plasmin formation were calculated by dividing the slope determined from the linear portion of the plot of A 405 versus time squared by the specific activity of plasmin (0.017 optical density units s Ϫ1 M Ϫ1 ), which was determined in a separate experiment (5).
Effect of CPB or TAFIa on (DD)E Degradation during t-PA-mediated Pg Activation-Plasmin degrades (DD)E to DD and fragment E (8,9). Because exposure of (DD)E to CPB or TAFIa reduces its ability to potentiate Pg activation, we examined whether CPB or TAFIa indirectly attenuates (DD)E degradation during t-PA-mediated Pg activation. 4 M (DD)E that had been exposed to 20 nM CPB for various intervals or to 20 nM TAFIa for 40 min was compared with 4 M native (DD)E in terms of its ability to stimulate 0.4 M Glu-Pg activation by 1 nM t-PA. Under these conditions, (DD)E concentrations were sufficient to monitor its degradation by PAGE analysis. After 1 h of incubation, the reaction was made to 1 M VFKCK and 10 nM PPACK to inhibit plasmin and t-PA, respectively, and a 4-l aliquot of the reaction solution was then subjected to PAGE analysis under nondenaturing conditions on 4 -15% gradient gels, on 10% gels in the presence of SDS, and on 15% gels in the presence of SDS and ␤-mercaptoethanol. Gels were stained with Fast Stain (Zoion Research, Shrewsbury, MA), and bands were quantified using an ImageMaster Video Documentation System. For amino acid sequencing, protein bands were transferred onto polyvinylidene difluoride membranes and stained with Ponceau S. Appropriate bands were cut from the membrane and subjected to amino-terminal sequence analysis, which was performed by Biotechnology Service Center (University of Toronto).
Determination of the Release of Free Lysine and Arginine from (DD)E by CPB or TAFIa-To determine whether CPB or TAFIa releases lysine and/or arginine residues from (DD)E, (DD)E was incubated with 20 nM CPB or TAFIa for intervals up to 60 min, and free lysine and arginine were measured using methods similar to those described by Wang et al. (19). After dialysis of (DD)E into 50 mM HEPES, 150 mM NaCl, pH 7.4, an 8 M (DD)E solution was incubated with 20 nM CPB for 60 min at 22°C. At various times, 100-l aliquots were removed, and 10-l of 40 M CPI was added to inactivate the CPB or TAFIa. The samples were deprotonated by bringing the solution to 0.2 M with perchloric acid followed by centrifuging at 12,000 ϫ g for 5 min. After the supernatants were neutralized with potassium hydroxide, the samples were placed on ice, and insoluble potassium perchlorate was removed by centrifugation at 12,000 ϫ g for 5 min. The concentration of lysine and arginine in the supernatants was determined enzymatically by methods established by Nakatani et al. (29) and Gaede et al. (30), respectively. For lysine determination, 50 l of supernatant was added to 40 l of 0.5 mM NADH (Roche Diagnostics, Laval, Canada) and 2.5 mM ␣-ketoglutaric acid (Roche Diagnostics) in 50 mM HEPES, 150 mM NaCl, pH 7.0 (HBS). Dehydration of NADH was initiated by the addition of 0.33 units of saccharopine dehydrogenase (Sigma) suspended in 10 l of HBS. Decreases in fluorescence intensity were monitored over 25 min in a Spectra Max, Gemini XS fluorescent plate reader (Molecular Devices, Sunnyvale, CA), with excitation and emission wavelengths set to 340 and 450 nm, respectively, and fitted with a 435-nm emission cut-off filter. The concentration of lysine was calculated based on a standard curve generated by plotting changes in fluorescence intensity produced by known concentrations of L-lysine (Sigma). Determination of free arginine was accomplished in the same manner using pyruvate (Roche Diagnostics) in place of ␣-ketoglutaric acid and 0.5 units of octopine dehydrogenase (Sigma) in place of saccharopine dehydrogenase. Standard curves for free arginine determination were generated by plotting changes in fluorescence intensity produced by known concentrations of L-arginine (Sigma).
Effect of (DD)E Exposure to TAFIa or CPB on the Kinetics of (DD)Estimulated Glu-and Lys-Pg Activation by t-PA-To determine the effects of exposure of (DD)E to TAFIa or CPB on the kinetics of (DD)Estimulated Pg activation, kinetic parameters for Glu-and Lys-Pg activation by t-PA, measured in the presence of TAFIa-or CPB-exposed (DD)E, were compared with those obtained with native (DD)E or in the absence of a cofactor. After exposure of 8 M (DD)E to 20 nM TAFIa or CPB for 40 min at 22°C, the reaction was terminated with CPI. Different concentrations of CPI were used to maintain a concentration of CPI in the Pg activation assay between 0.2 and 0.8 M. This range of CPI concentrations was maintained to ensure CPI levels were well above the K i for inhibiting CPB (31) but low enough so as not to inhibit plasmin (5 M CPI causes a 7% reduction in the activity of 0.01 M plasmin). 0.8 M CPI has no measurable effect on the rate of plasmin formation (data not shown). Glu-or Lys-Pg, in concentrations ranging from 0 to 16 M, was incubated with 0.1 nM t-PA in the absence or presence of native (DD)E or TAFIa-or CPB-exposed (DD)E at concentrations up to 6 M. Plasmin formation was then monitored using the plasmin-directed substrate S-2251 and rates of Pg activation (), determined as described above, were divided by the concentration of activator (Ј) and subjected to nonlinear regression analysis (Table Curve; Jandel Scientific, San Rafael, CA) according to the Michaelis-Menten is the equilibrium concentration of free substrate, calculated for each input concentration of Pg and cofactor as described (3). Free Glu-and Lys-Pg concentrations were calculated based on their affinities for native (DD)E (K d values of 5.5 and 0.09, respectively) or CPB-exposed (DD)E (K d values of 11 and 0.35 M, respectively), which were determined as described below. Michaelis-Menten analysis yielded the rate constant (k cat ) and Michaelis constant (K m ). -Fold stimulation of Pg activation was calculated by dividing the catalytic efficiency (k cat /K m ) obtained in the presence of saturating concentrations of native or TAFIa-or CPB-exposed (DD)E with that calculated in its absence. Cofactor concentrations were considered saturating when the catalytic efficiency reached a maximum. For comparison purposes, the effect of fibrin and Fg on Glu-and Lys-Pg activation by t-PA was determined as described previously (11).
Effect of (DD)E Exposure to TAFIa or CPB on Its Affinity for t-PA, Glu-Pg, or Lys-Pg-The affinities of native or TAFIa-or CPB-exposed (DD)E for t-PA, Glu-Pg, or Lys-Pg were quantified using right angle light-scattering spectroscopy in a LS50B luminescence spectrometer (PerkinElmer Life Sciences) (5,11). 15-or 20-l aliquots of active site-blocked t-PA, Glu-Pg, or Lys-Pg (at concentrations of 6, 100, and 10 M, respectively) were added to 2 ml of a 0.1 M solution of native or TAFIa-or CPB-exposed (DD)E in a 3-ml quartz cuvette thermostated at 22°C with stirring. The solution was excited at 400 or 440 nm, and emission intensities were measured at the same wavelength. Excitation and emission slit widths were set to either 8 or 12 nm, and a 1% attenuation emission filter was used. Changes in scattering intensity were continuously monitored in time drive with the interval time set at 2 s and the response time set at 3 s. Scattering intensities after each addition of ligand were obtained from the time drive profile by averaging scattering intensities observed over a period of at least 100 s. Control titrations were performed to determine the light scattering intensity of t-PA, Glu-Pg, or Lys-Pg alone. The emission intensity (I) of the incident beam after each addition of ligand was corrected for changes due to ligand scattering and dilution. Corrected values were compared with the emission intensity of the target alone (I o ). When TAFIa-or CPB-exposed (DD)E was used, scattering intensities were corrected for scattering due to the presence of CPI and TAFIa or CPB. These data, together with the total ligand concentration (L o ), were fit by nonlinear regression analysis (Table Curve, Jandel Scientific) to Equation 1 (5,11,32), where L o is the total concentration of ligand added, P o is the concentration of target protein, K d is the dissociation constant, n is the stoichiometry, and ␣ is the maximum change in emission intensity. To confirm the binding parameters obtained in this fashion, reverse titrations also were performed wherein 0.1 M active site-blocked t-PA, Glu-Pg, or Lys-Pg was titrated with native or TAFIa-or CPB-exposed (DD)E.

RESULTS
Effect of CPB or TAFIa on the Ability of (DD)E to Stimulate t-PA-mediated Pg Activation-To determine whether CPB or TAFIa treatment of (DD)E modulates its ability to stimulate Pg activation by t-PA, (DD)E was incubated with CPB or TAFIa for intervals up to 40 min. After inhibiting CPB or TAFIa with CPI, the stimulatory activity of CPB-or TAFIa-treated (DD)E was then compared with that of the starting material. When t-PA-mediated activation of Glu-Pg is monitored, plasmin formation is decreased with increasing exposure of (DD)E to CPB (Fig. 1, inset) or TAFIa (data not shown). From these data, rates of plasmin formation were calculated (Fig. 1). After a 40-min exposure of (DD)E to CPB or TAFIa, the rate of Glu-Pg activation in the presence of CPB-treated (DD)E is similar to that in the absence of a cofactor. Analogous results are obtained when Lys-Pg is substituted for Glu-Pg. Although rates of plasmin formation are higher with Lys-Pg than with Glu-Pg, 40-min exposure of (DD)E to CPB or TAFIa almost completely abolishes its stimulatory activity (Fig. 2).
Plasmin can degrade (DD)E into DD and fragment E (8,9). Because DD and fragment E are less potent stimulators of Pg activation than intact (DD)E (12), this phenomenon could influence the kinetics of (DD)E-stimulated Pg activation. However, when analyzed by PAGE, no detectable (DD)E degradation is observed during the first 10 min of Pg activation (data not shown), the time over which the initial rates of plasmin formation are calculated. These observations suggest that degradation of (DD)E does not influence the kinetics of Pg activation in the presence of (DD)E.
Effect of CPB or TAFIa on the Degree to Which (DD)E Is Degraded during t-PA-mediated Pg Activation- Fig. 3, which illustrates the formation of (DD)E from fibrin and its subsequent degradation by plasmin, helps to describe the gels presented in Fig. 4. (DD)E, which is formed when plasmin degrades two-stranded fibrin protofibrils between adjacent D and E domains, consists of two cross-linked D domains noncovalently associated with an E domain (Fig. 3). The E domain remains bound to DD provided that the amino-terminal portion of at least one ␣ and ␤ chain within the E domain (␣ E and ␤ E , respectively) remains intact (25,33). Plasmin first cleaves ␤ E at Lys 53 and then ␣ E at Arg 19 . The E moiety is characterized by changes in molecular weight that occur upon ␤ E cleavage (33,34). Fragment E 1 , which has a molecular mass of ϳ60 kDa, has the Lys 53 -Lys 54 peptide bonds of both ␤ E chains intact, whereas fragment E 2 , with a molecular mass of 55 kDa, has one ␤ E chain cleaved at Lys 53 . Both fragments E 1 and E 2 remain associated with DD. In contrast, fragment E 3 , which has a molecular mass of ϳ50 kDa, no longer binds DD, because both ␤ E chains are cleaved at Lys 53 .
When (DD)E is exposed to plasmin, it is degraded to DD and fragment E (8,9). Since TAFIa attenuates fibrin degradation by limiting plasmin formation, we explored the possibility that CPB or TAFIa modulates (DD)E degradation (15,16). (DD)E was incubated with CPB for 40 min. At intervals, aliquots were removed, and CPB was inhibited with CPI. Each (DD)E aliquot was incubated with t-PA and Pg for 1 h to promote Pg activation. Using PAGE analysis, the extent to which the various (DD)E samples were degraded was examined. Under nondenaturing conditions, native (DD)E migrates as a single band (Fig.  4A, lanes 1 and 9). When Pg is activated by t-PA in the presence of native (DD)E, (DD)E is degraded to two lower molecular weight species, corresponding to DD and fragment E, respectively (Fig. 4A, lane 2). In contrast, when (DD)E is preincu-bated with CPB for increasing amounts of time before incubation with t-PA and Pg (Fig. 4A, lanes 3-8), there is a noticeable decrease in the mobility of fragment E and a progressive increase in the amount of intact (DD)E. Under denaturing conditions, native (DD)E migrates as two bands corresponding to DD and fragment E, respectively (Fig. 4B, lane 1). Fragment E from (DD)E is a single species corresponding to fragment E 1 , characterized by ␤ chains with amino-terminal Gly residues at position 15 (33). Following a 1-h incubation in the presence of t-PA and Pg, DD mobility is unchanged, but fragment E 1 is degraded to E 3 (Fig. 4B, lane 2), a species characterized by amino-terminal Lys residues at position 54 on both ␤ chains (33). With increasing exposure of (DD)E to CPB, there is less degradation of fragment E, until E 1 is the only fragment E species in (DD)E (Fig. 4B, lanes 3-8). Under reducing and denaturing conditions, the six chains of native (DD)E were identified based on their apparent molecular weight and amino-terminal sequence analysis. These are the ␣ and ␤ chains and ␥-␥ dimer from DD (␣ D , ␤ D , and ␥-␥ D , respectively) and ␣, ␤, and ␥ chains from fragment E (␣ E , ␤ E , and ␥ E , respectively) (Fig. 4C, lanes 1 and 9). When native (DD)E stimulates Pg activation for 1 h, the ␤ E chain is degraded, and, based on amino-terminal sequence analysis, three amino acids are removed from the amino-terminal portion of the ␣ E chain, although there is no noticeable shift in mobility of the ␣ E chain (Fig. 4C, lane 2). All other chains remain intact. These data are consistent with previously reported plasmin cleavage sites (33). When CPB-exposed (DD)E is used to stimulate Pg activation, the ␤ E chain is not degraded (Fig. 4C, lanes 3-8). The ␣ E chain also remains intact, as determined by its amino-terminal sequence.
The amount of ␤ E chain in each lane in Fig. 4C was quantified by measuring its optical density and comparing this value to the total optical density of all chains in the same lane. As illustrated in Fig. 5, the extent of ␤ E chain degradation is inversely related to the rate of plasmin formation measured prior to PAGE analysis (r ϭ Ϫ0.986).
To determine whether TAFIa-mediated attenuation of plasmin formation also limits the degradation of (DD)E, (DD)E was incubated with 20 nM TAFIa for 40 min. At the end of the incubation period, CPI was added to stop the reaction. Like the results with CPB, when TAFIa-exposed (DD)E is used to stimulate Pg activation, the ␤ E and ␣ E chains are not degraded, and

FIG. 1. Effect of CPB or TAFIa on the rate of (DD)E-stimulated activation of Glu-Pg by t-PA. 8 M (DD)E was
incubated with 20 nM CPB or TAFIa for 40 min at 22°C. At the indicated intervals, aliquots were removed, and CPB or TAFIa was inactivated with CPI. CPB-or TAFIa-exposed (DD)E (0.4 M) was incubated with 0.4 M Glu-Pg and 0.1 nM t-PA for 60 min at 37°C, and plasmin formation was monitored at 405 nm using the plasmin-directed substrate S-2251. The inset shows the raw data for plasmin generation in the presence of CPB-exposed (DD)E. From these data, rates of plasmin formation were calculated. Each point represents the mean Ϯ S.E. of at least three experiments. The rate of Pg activation in the presence of (DD)E incubated for 40 min with CPB (ⅷ) or TAFIa (f) is similar to that measured in the absence of (DD)E.

FIG. 2. The effect of CPB or TAFIa on the rate of (DD)E-stimulated activation of Glu-or Lys-Pg by t-PA.
To determine whether exposure of (DD)E to CPB or TAFIa decreases its ability to stimulate the activation of Lys-Pg to the same extent as Glu-Pg, 8 M (DD)E was incubated with 20 nM CPB or TAFIa for 40 min at 22°C. After the addition of CPI, 0.4 M CPB-or TAFIa-exposed (DD)E was incubated with 0.4 M Glu-Pg or 0.1 M Lys-Pg and 0.1 nM t-PA, and rates of plasmin formation were determined. The rates obtained in the presence of 0.4 M native (DD)E were used to calculate the relative rates for CPBor TAFIa-exposed (DD)E. Each bar represents the mean Ϯ S.E. of at least three experiments. Exposure of (DD)E to CPB or TAFIa reduces its ability to stimulate Glu-or Lys-Pg activation by t-PA to almost background levels.
the structural integrity of (DD)E is preserved (data not shown).
CPB-or TAFIa-mediated Release of Lysine and Arginine from (DD)E-To explore the mechanism by which CPB or TA-FIa attenuates the stimulatory activity of (DD)E, we determined whether CPB or TAFIa induces the release of free lysine and arginine residues from (DD)E. Under the conditions outlined under "Methods," CPB or TAFIa causes the release of approximately 4 mol of lysine/mol of (DD)E within 10 and 2 min, respectively (Fig. 6). By 60 min, an additional 4 mol of lysine/mol of (DD)E are released by both enzymes. The biphasic lysine release from (DD)E is not the result of reduction in the activity of CPB, because CPB activity against hippuryl-L-arginine remains constant over the 60-min incubation period. Although the activity of TAFIa against hippuryl-L-arginine is 25% lower at 60 min than it is to start, there is no detectable change in TAFIa activity at 10 min, a point well beyond the transition from slow to rapid lysine release. Thus, the carboxylterminal lysine residues on (DD)E exhibit different susceptibilities to CPB or TAFIa. Neither CPB nor TAFIa causes the release of arginine from (DD)E (Fig. 6).

Effect of (DD)E Exposure to TAFIa or CPB on the Kinetics of (DD)E-Stimulated Glu-and Lys-Pg Activation by t-PA-The
kinetic parameters for (DD)E-stimulated Glu-or Lys-Pg activation by t-PA were measured in the absence or presence of native (DD)E, fibrin or Fg. These results were then compared with those obtained with TAFIa-or CPB-exposed (DD)E. Gluor Lys-Pg was activated with a constant amount of t-PA in the presence of varying concentrations of Pg and cofactor. Data obtained with each cofactor concentration were fit to the Michaelis-Menten equation by nonlinear regression to determine the values of k cat and K m (Fig. 7). Cofactor concentrations were considered saturating when the catalytic efficiency (k cat / K m ) reached a maximum. Maximum catalytic efficiencies were achieved with 0.4 M native (DD)E and 5 M (DD)E previously exposed to TAFIa or CPB.
As summarized in Table I, native (DD)E is as potent as fibrin at potentiating both Glu-and Lys-Pg activation by t-PA, stim-

FIG. 3. Model of (DD)E formation and degradation. Thrombin releases fibrinopeptide A (A␣ 1-16 ) and fibrinopeptide B (B␤ 1-14 ) from
Fg, thereby generating fibrin monomers. These aggregate to form twostranded protofibrils that are arranged in a half-staggered fashion, whereby the E domain of a fibrin monomer of one strand is noncovalently associated with D domains of two adjacent fibrin monomers on the opposite strand. Factor XIIIa catalyzes the cross-linking of ␥ chains of adjacent D domains. Plasmin cleaves the ␣, ␤, and ␥ chains between E and D domains to yield (DD)E 1 , wherein fragment E 1 remains noncovalently associated with DD via two "knob and hole" interactions on each D domain. These interactions only occur when the amino termini of the ␣ and ␤ chains of fragment E are intact. Exposure of (DD)E 1 to plasmin results in progressive cleavage of the ␤ and ␣ chains of fragment E. The E moiety is characterized by cleavage of the Lys 53 -Lys 54 peptide bonds on the ␤ chains. Fragment E 2 , which has one ␤ chain cleaved, remains noncovalently associated with DD via one knob and hole interaction. In contrast, fragment E 3 , which has both ␤ chains cleaved at position 53, no longer interacts with DD. Cleavage of both ␣ chains at position 19 within fragment E occurs subsequent to cleavage of the ␤ chains. , denaturing conditions on a 10% gel (B), or denaturing and reducing conditions on a 15% gel (C). Lane numbers are indicated above each gel, and the time (DD)E was exposed to CPB prior to Pg activation is indicated above the gel marked A. Native (DD)E is shown in lanes 1 and 9, and molecular weight markers, when present, are shown in lane 10. Native (DD)E consists only of DD and fragment E 1 . When native (DD)E is used to stimulate Pg activation, the ␤ E chain is completely degraded, fragment E 1 is degraded to E 3 , and the E moiety no longer remains associated with DD. In contrast, when CPB-exposed (DD)E is used, the ␤ E chain is not degraded, and fragment E 1 remains intact and associated with DD. Thus, by limiting plasmin formation, CPB indirectly preserves the structural integrity of (DD)E. ulating these reactions 360-and 450-fold, respectively. In contrast, after CPB or TAFIa exposure, (DD)E-stimulation of both Glu-and Lys-Pg activation is reduced by 96% (Table I). These reductions reflect decreases in k cat values and increases in K m values. The k cat values for Glu-and Lys-Pg activation with CPB-exposed (DD)E are 3-and 2-fold lower, respectively, than those with native (DD)E, whereas K m values for these reactions are increased 10-and 12-fold, respectively. Consequently, TA-FIa-or CPB-treated (DD)E stimulates Glu-and Lys-Pg activation only 12-and 18-fold, respectively, stimulatory activity similar to that of Fg.
Effect of TAFIa or CPB on the Binding of t-PA, Glu-Pg, and Lys-Pg to (DD)E-To determine whether the reduced stimulatory activity of TAFIa-or CPB-treated (DD)E reflects changes in its affinity for t-PA and/or Pg, light scattering spectroscopy was used to compare the affinities of t-PA or Pg for native (DD)E with those for TAFIa-or CPB-exposed (DD)E. The relative scatter plots for the interactions of t-PA or Glu-Pg with (DD)E before and after CPB treatment are illustrated in Fig. 8,  A and B, respectively. Under the conditions outlined under "Methods," the scattering intensity of 0.1 M (DD)E is 15 (I o ). At saturating levels of t-PA, the maximum relative scattering intensity (I/I o ) in the presence of untreated (DD)E is 1.5 (Fig.  8A), a value in good agreement with a calculated maximum relative scattering intensity of 1.6 if the stoichiometry is 1:1 (5,35). The maximum I/I o when (DD)E was titrated with Glu-Pg was 2.1 (Fig. 8B), a value similar to a predicted I/I o of 1.9 for a 1:1 substrate interaction with (DD)E (5,35). Data were fit to Equation 1 by nonlinear regression analysis to determine dissociation constants, and these results are summarized in Table  II. t-PA binds to native (DD)E with a K d of 0.04 M. In contrast, t-PA binds to CPB-exposed (DD)E with a K d of 6.5 M, an affinity 160-fold lower than that of the activator for native (DD)E. Glu-Pg binds to native and CPB-exposed (DD)E with K d values of 5.5 and 11 M, respectively. Thus, CPB treatment produces only a 2-fold decrease in the affinity of (DD)E for Glu-Pg. Similar results are obtained with Lys-Pg, which binds to CPB-exposed (DD)E with a 4-fold lower affinity than that for native (DD)E (K d ϭ 0.35 and 0.09 M, respectively). The affinities of TAFIa-exposed (DD)E for t-PA, Glu-Pg, or Lys-Pg are similar to those of CPB-exposed (DD)E (Table II). DISCUSSION (DD)E, a soluble degradation product of cross-linked fibrin, compromises the fibrin specificity of t-PA by promoting systemic activation of Pg (5,8,12). The stimulatory activity of (DD)E, like that of fibrin, reflects its capacity to bind t-PA and Pg (5,10,12). In contrast to its predominantly finger-mediated interaction with fibrin, t-PA binds to (DD)E via its second kringle domain (5,11). Although Pg binding also is kringle-dependent, Pg does not compete with t-PA for (DD)E binding, indicating that Pg and t-PA bind to distinct lysine residues on (DD)E (5). By releasing carboxyl-terminal lysine residues that promote Pg binding to fibrin, TAFIa, a CPB-like enzyme, attenuates fibrin degradation (15,17). In this study, we explored FIG. 5. Correlation between ␤ E chain degradation and the ability of (DD)E to stimulate Pg activation. The relative amounts of ␤ E in each lane of Fig. 4C (f), determined by densitometry, are plotted against the time that (DD)E was exposed to CPB prior to Pg activation. The rate of Pg activation by CPB-exposed (DD)E (ⅷ), measured prior to PAGE analysis, is plotted against time of CPB exposure on the secondary y axis. Both of these values are compared with those obtained with native (DD)E. These data demonstrate that the integrity of the ␤ E chain is inversely correlated (r ϭ Ϫ0.986) with the extent to which (DD)E stimulates Pg activation.
FIG. 6. CPB-or TAFIa-mediated lysine and arginine release from (DD)E. 8 M (DD)E was incubated with 20 nM CPB or TAFIa for various times at 22°C. After the addition of CPI, the amount of free lysine or arginine release was determined by monitoring dehydration of NADH using saccharopine or octopine dehydrogenase, respectively. The total concentration of lysine or arginine was divided by the concentration of (DD)E to calculate mol of free amino acid released/mol of (DD)E. CPB (ⅷ) or TAFIa (f) causes the release of 8 mol of lysine/mol of (DD)E. Under the conditions employed here, ϳ4 mol of lysine are released within the first 10 and 2 min of CPB and TAFIa incubation, respectively. The remaining 4 mol of lysine are released within 60 min of incubation with either CPB or TAFIa. Neither CPB (⅜) nor TAFIa (Ⅺ) releases arginine residues from (DD)E. the possibility that TAFIa modulates the stimulatory activity of (DD)E via a similar mechanism.
(DD)E is as potent as fibrin at stimulating t-PA-mediated Pg activation. When (DD)E is incubated with TAFIa or CPB, its ability to stimulate the activation of either Glu-or Lys-Pg by t-PA is reduced by 96% and becomes comparable with that TABLE I Kinetic parameters for Glu-and Lys-Pg activation in the absence of a cofactor, in the presence of (DD)E before and after its exposure to CPB or TAFIa, or in the presence of fibrin or Fg All kinetic parameters are presented as the mean Ϯ S.E. of at least three experiments, except those for TAFIa-exposed (DD)E, which are in parentheses and reported as the mean of two experiments.   Table II. Whereas exposure of (DD)E to CPB causes a modest reduction in its affinity for Glu-Pg, CPB markedly reduces the affinity of (DD)E for t-PA.
of Fg. Reduced stimulation by (DD)E reflects both a decrease in k cat and an increase in K m for both Glu-and Lys-Pg. The larger effect is on K m , however, suggesting that exposure of (DD)E to TAFIa or CPB serves to destabilize the ternary t-PA-plasminogen-(DD)E complex.
Lysine residues are released in parallel with the loss of stimulatory activity that occurs as (DD)E is incubated with TAFIa or CPB. A total of 8 mol of lysine/mol of (DD)E is generated. Based on the molecular weight and amino-terminal sequence of the six individual chains within (DD)E as well as the known plasmin cleavage sites on fibrin, the 8 lysine residues released correspond to the 6 carboxyl-terminal lysine residues on the ␣, ␤, and ␥ chains of fragment E at positions 78, 133, and 62, respectively, as well as the 2 carboxyl-terminal lysine residues on the ␣ chains of each D-domain at position 206 (33,34,36). Under the conditions employed in this study, most of the stimulatory activity of (DD)E is lost after 2-or 10-min incubation with TAFIa or CPB, respectively, times at which ϳ4 mol of lysine/mol of (DD)E are released. These findings suggest that some lysine residues on (DD)E are more susceptible to TAFIa-or CPB-mediated release than others and that those most susceptible to release are the lysine residues important for formation of the ternary t-PA-plasminogen-(DD)E complex.
Because plasmin generation in the presence of TAFIa-or CPB-exposed (DD)E is less than that with native (DD)E, TAFIa or CPB indirectly preserves the structural integrity of (DD)E. This may be of physiologic relevance, because DD binds to fibrin with high affinity and inhibits its polymerization, whereas intact (DD)E has no effect on fibrin polymerization (37)(38)(39). Limiting DD formation may promote clot stability, because fibrin clots are dynamic structures that undergo continuous formation and degradation. Thus, the anti-fibrinolytic properties of TAFIa may include clot stabilization by the prevention of (DD)E degradation to DD and fragment E.
To explore the mechanism by which TAFIa or CPB eliminates the cofactor activity of (DD)E, we compared the affinities of t-PA and Pg for TAFIa-or CPB-treated (DD)E with those for native (DD)E. Exposure of (DD)E to TAFIa or CPB reduces its affinity for Glu-and Lys-Pg 2-and 4-fold, respectively. In contrast, the affinity of (DD)E for t-PA is reduced 160-fold after TAFIa or CPB exposure. The fact that TAFIa or CPB exposure lowers the affinity of (DD)E for t-PA more than it reduces its affinity for Glu-or Lys-Pg, suggests that TAFIa or CPB modulates the stimulatory activity of (DD)E primarily by decreasing its affinity for the activator rather than for plasminogen.
Because TAFIa-or CPB-exposed (DD)E has little affinity for t-PA, it is a poor stimulator of Glu-or Lys-Pg activation by t-PA. This observation is consistent with our previous work with the Pg activator from the saliva of the vampire bat (b-PA). Lacking a lysine-binding kringle, b-PA does not bind (DD)E (5).
Consequently, (DD)E is a poor stimulator of Pg activation by b-PA (5,40). These data support the concept that only Pg activators that bind to (DD)E are significantly potentiated by this fragment.
Although it does not bind to (DD)E, b-PA binds fibrin with high affinity via its finger domain and is stimulated by fibrin to the same extent as t-PA (5,41). Because (DD)E and fibrin stimulate t-PA to a similar extent, whereas b-PA is potentiated only by fibrin, t-PA is less fibrin-specific than b-PA. Thus, (DD)E compromises the fibrin specificity of t-PA. Our current findings raise the possibility that by decreasing the capacity of (DD)E to bind t-PA, TAFIa enhances the fibrin specificity of t-PA.
Upon exposure to TAFIa or CPB, the affinity of (DD)E for Glu-or Lys-Pg is only modestly reduced, suggesting that Pg binds to internal lysines as well as carboxyl-terminal lysine residues. In contrast, our data suggest that t-PA predominantly binds to carboxyl-terminal lysine residues on (DD)E because the affinity of t-PA for (DD)E is markedly reduced when (DD)E is exposed to TAFIa or CPB. These findings are consistent with the previously reported noncompetitive binding of t-PA and Pg to (DD)E (5).
Like their effect on (DD)E, TAFIa and CPB also release lysine residues from fibrin partially degraded by plasmin (19,42). This results in a reduction in the affinity of plasminexposed fibrin for Glu-or Lys-Pg and blocks the accumulation of fluorescently labeled Pg on fibrin (16,43). TAFIa also prevents the conversion of Glu-Pg to Lys-Pg, an early event in the course of t-PA-mediated clot lysis that serves as a positive feedback mechanism because Lys-Pg, which has higher affinity for fibrin than Glu-Pg, is more readily activated by t-PA (18,19). Although the effects of TAFIa on the affinity of the activator for fibrin have yet to be investigated, CPB blocks the 2-3fold increase in t-PA binding that occurs when fibrin is exposed to plasmin, suggesting that this increase reflects kringle-dependent binding of t-PA to newly exposed carboxyl-terminal lysine residues (42). The inhibitory effect of TAFIa or CPB on these positive feedback events results in a 3-4-fold prolongation of the rate of t-PA-mediated fibrinolysis (15,26).
Our results with (DD)E have similarities to those with fibrin. TAFIa or CPB causes the release of free lysine residues from (DD)E, reduces the affinity of the substrate and activator for the cofactor, and attenuates cofactor degradation. However, other features distinguish the effects of TAFIa or CPB on (DD)E from those on fibrin. Although fibrin degradation is attenuated by TAFIa or CPB, it is not inhibited, suggesting that even in the presence of these enzymes, fibrin remains a competent stimulator of Pg activation by t-PA (26). Furthermore, the inhibitory effects of TAFIa on fibrinolysis can be overcome by substituting Lys-Pg for Glu-Pg (18,19). In contrast, exposure of (DD)E to TAFIa or CPB produces a 96% reduction in its ability to stimulate the activation of either Gluor Lys-Pg by t-PA. These differences between fibrin and (DD)E are explained by the observation that TAFIa or CPB reduces the affinity of t-PA for (DD)E more than its affinity for fibrin. Whereas t-PA maintains high affinity binding to intact or partially degraded fibrin in the presence of CPB (42,44), TAFIa or CPB almost abolishes t-PA binding to (DD)E.
(DD)E compromises the fibrin specificity of t-PA because it binds t-PA with high affinity and stimulates Pg activation to the same extent as fibrin. By reducing the affinity of (DD)E for t-PA, TAFIa converts (DD)E from a fibrin-like stimulator to one that has minimal stimulatory activity, much like Fg. Thus, our data suggest a novel role for TAFIa. By reducing the stimulatory activity of (DD)E and attenuating systemic Pg activation, TAFIa may enhance the fibrin specificity of t-PA. Furthermore, a Values for native and CPB-exposed (DD)E are presented as the means Ϯ S.E. of at least three titrations.
b -Fold decrease in affinity is calculated by dividing the affinity of native (DD)E for the indicated ligand by the affinity of CPB-exposed (DD)E for the same ligand.
c Values in parentheses are for TAFIa-exposed (DD)E and are presented as the mean of two titrations. since DD can impair fibrin polymerization, limiting (DD)E degradation may augment the anti-fibrinolytic properties of TAFIa.