Characterization of plasmin-mediated activation of plasma procarboxypeptidase B. Modulation by glycosaminoglycans.

Plasma carboxypeptidase B (PCB) is an exopeptidase that exerts an antifibrinolytic effect by releasing C-terminal Lys and Arg residues from partially degraded fibrin. PCB is produced in plasma via limited proteolysis of the zymogen, pro-PCB. In this report, we show that the K(m) (55 nM) for plasmin-catalyzed activation of pro-PCB is similar to the plasma concentration of pro-PCB (50-70 nM), whereas the K(m) for the thrombin- or thrombin:thrombomodulin-catalyzed reaction is 10-40-fold higher than the pro-PCB level in plasma. Additionally, tissue-type plasminogen activator triggers activation of pro-PCB in blood plasma in a reaction that is stimulated by a neutralizing antibody versus alpha(2)-antiplasmin. Together, these results show that plasmin-mediated activation of pro-PCB can occur in blood plasma. Heparin (UH) and other anionic glycosaminoglycans stimulate pro-PCB activation by plasmin but not by thrombin or thrombin:thrombomodulin. Pro-PCB is a more favorable substrate for plasmin in the presence of UH (16-fold increase in k(cat)/K(m)). UH also stabilizes PCB against spontaneous inactivation. The presence of UH in clots prepared with prothrombin-deficient plasma delays tissue-type plasminogen activator-triggered lysis; this effect of UH on clot lysis is blocked by a PCB inhibitor from potato tubers. These results show that UH accelerates plasmin-catalyzed activation of pro-PCB in plasma and PCB, in turn, stabilizes fibrin against fibrinolysis. We propose that glycosaminoglycans in the subendothelial extracellular matrix serve to augment the levels of PCB activity thereby stabilizing blood clots at sites where there is a breach in the integrity of the vasculature.

The exopeptidase activity of plasma carboxypeptidase B (PCB) 1,2 is postulated to play a key regulatory role in the fibrinolytic cascade (1,2). The zymogen of PCB, pro-PCB, is activated by limited proteolysis to expose carboxypeptidase activity directed at C-terminal Lys and Arg residues (3,4). Partially degraded fibrin is decorated with C-terminal basic residues that serve to accelerate fibrinolysis by providing binding sites for tissue-type plasminogen activator (t-PA), plasminogen (Plg), and plasmin (5). The ability of PCB to release C-terminal Lys and Arg residues from partially degraded fibrin thereby exerts an antifibrinolytic effect.
Thrombin was shown to activate pro-PCB albeit with low catalytic efficiency (4,6). More recently, Bajzar et al. (7) demonstrated that thrombin-catalyzed pro-PCB activation is stimulated 1250-fold by thrombomodulin (TM), a membrane-associated thrombin-binding protein. Steady-state kinetics revealed that TM increased the turnover rate (k cat ) of the Michaelis-Menten complex, but the Michaelis constant (K m ) was essentially unchanged. Trypsin and plasmin were also reported to catalyze pro-PCB activation (4).
Plasmin plays a pivotal role in clot dissolution by catalyzing the proteolytic cleavage of fibrin. Plasmin-mediated fibrinolysis is subject to numerous regulatory controls to deter the bleeding diathesis that would accompany unrestricted clot lysis. In this report, we characterize a novel regulatory pathway that acts to suppress fibrinolysis. Activation of pro-PCB by plasmin to generate PCB is a negative feedback mechanism in which plasmin activity serves to dampen its familiar profibrinolytic effect. Furthermore, glycosaminoglycans (GAGs) were shown to increase markedly the rate of plasmin-mediated pro-PCB activation and stabilize PCB activity against spontaneous inactivation. A breach in the integrity of the vasculature will serve to expose GAGs that are present in the extracellular matrix (8). Hence, the GAG-mediated effects that augment the levels of PCB activity thus inhibiting fibrinolysis could possibly stabilize the clot at sites of vascular injury.

EXPERIMENTAL PROCEDURES
Materials-Pro-PCB, plasmin, and rabbit lung TM were purchased from Hematologic Technologies, Inc. (Essex Junction, VT). The pro-PCB concentration was derived using A 280 nm ϭ 1.49 (mg/ml) Ϫ1 and an apparent relative molecular weight of 60,000 (4,6). Sheep anti-pro-PCB IgG was from Affinity Biologicals Inc. (Hamilton, Ontario). Recombinant t-PA (Activase) was from Genentech Inc. (South San Francisco, CA). Batroxobin, goat anti-human Plg IgG, and mouse anti-␣ 2 -antiplasmin (␣ 2 AP) monoclonal IgG were from American Diagnostica, Inc. (Greenwich, CT). Human histidine-rich glycoprotein was purified from fresh citrated plasma as described elsewhere (9). Potato carboxypeptidase inhibitor (PCI), UH from porcine intestinal mucosa (grade II; 167 USP unit/mg), low molecular weight heparin (M r 3000), heparan sulfate (i.e. heparin monosulfate), chondroitin sulfate A, chondroitin sulfate B (i.e. dermatan sulfate), chondroitin sulfate C, dextran sulfates (M r 5000 and 8000) and keratan sulfate were purchased from Sigma. The con-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Activation of Pro-PCB-Pro-PCB (50 nM) was added to 50 mM HEPES, pH 7.4, 0.15 M NaCl, and 0.1% PEG-8000 (HBSP, HEPESbuffered saline with 0.1% PEG 8000) containing CaCl 2 (2 mM), FA-AR (720 M), and various concentrations of test agents (final volume ϭ 250 l) in PEG-20,000 precoated 96-well microtiter plates. The K m value of PCB for FA-AR is 360 M. Plasmin or thrombin (10 nM) was added to initiate pro-PCB activation. The reaction was performed at room temperature with continuous monitoring of PCB activity using a Thermo-Max plate reader (Molecular Devices, Menlo Park, CA). PCB activity was detected by a decrease in A 340 nm due to hydrolysis of FA-AR (13).
Western Blot Analysis of Pro-PCB Activation-Pro-PCB (50 nM) was added to HBSP containing 2 mM CaCl 2 . Other additions (where indicated) included 10 nM thrombin Ϯ 10 nM TM, 10 nM plasmin, and 10 -100 units/ml UH (final volume ϭ 200 l). The reactions were performed at room temperature. Twenty-l aliquots were removed at increasing times and quenched with concentrated Laemmli loading buffer (40 l) containing 2-mercaptoethanol. Samples were fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis using 12 or 10 -20% gradient Tris glycine SDS gels (Novex, San Diego, CA) and electrophoretically transferred to nitrocellulose membranes (Novex). Membranes were blocked with non-fat dry milk and sequentially treated with sheep anti-pro-PCB IgG and rabbit peroxidase-conjugated anti-sheep IgG (Organon Teknika Co., West Chester, PA). Pro-PCB bands were detected with the SuperSignal chemiluminescent substrate (Pierce).
Stability of PCB Activity-Pro-PCB (500 nM) was activated with 25 nM thrombin and 25 nM TM for 40 min at 25°C. The thrombin activity was subsequently quenched with 200 nM hirudin. The sample was diluted 5-fold in HBSP buffer with and without 200 units/ml UH. At various times, PCB activity was assayed with 720 M FA-AR using a Shimazu UV-2101PC spectrophotometer. An exponential fit of each data set was used to derive a first-order rate constant (k inact ). The half-life of inactivation (t1 ⁄2 ) was calculated from the relationship t ½ ϭ 0.693/k inact .
N-terminal Sequencing of the PCB Subunit-Pro-PCB (2 M) was activated with 500 nM plasmin or 500 nM thrombin ϩ 120 nM TM for 1 h at room temperature in HBSP containing 2 mM CaCl 2 . Fifty-l aliquots were removed and quenched with concentrated Laemmli loading buffer containing 2-mercaptoethanol. Samples were fractionated by SDS-polyacrylamide gel electrophoresis using a 12% Tris glycine SDS gel and electrophoretically transferred to a polyvinylidene difluoride membrane (Novex). The membrane was briefly stained with Coomassie Blue, and the 35-kDa bands were excised. The N-terminal sequences were determined with an Applied Biosystem Model 494 Protein Sequencer (Foster City, CA).
Assay for PCB-mediated Arg Release from Plasmin-Active-site blocked plasmin (FFR-Pln) was prepared by treating plasmin (10 M) with D-Phe-L-Phe-L-Arg-chloromethyl ketone (200 M) in HBSP buffer for 30 min at room temperature. FFR-Pln or plasmin (6 M) was incubated with PCB (200 nM) at 37°C in HBSP buffer. At increasing times, 50-l aliquots were removed, and the PCB activity was quenched with 25 l of Plummer's inhibitor (10 M). Samples were lyophilized to dryness, resuspended in 10 l of borate buffer (0.4 M, pH 10.2), and analyzed for Arg using a Hewlett-Packard Amino Quant II/M High Sensitivity System (which includes a Hewlett-Packard 1090 Liquid Chromatograph with a 2.1 ϫ 200 mm HP Amino Quant column, a Hewlett-Packard 1046A Programmable Fluorescence Detector, and an HPLC Chemstation). Mobile phase A was 20 mM sodium acetate buffer containing 0.018% triethylamine, adjusted to pH 7.2 with 1-2% acetic acid, and 0.3% tetrahydrofuran. Mobile phase B was 20% 0.1 M sodium acetate buffer, adjusted to pH 7.2 with 1-2% acetic acid, 40% acetonitrile, and 40% methanol. Primary amino acids in each sample (1-ml aliquots) were subjected to pre-column derivatization with o-phthalde-hyde and subsequently detected by fluorescence using excitation and emission wavelengths of 340 and 450 nm, respectively. Arg eluted in this system at approximately 7.0 min. The amounts of released Arg were interpolated from Arg standard curves.
Impact of PCB on Plasmin-mediated Pro-PCB Activation-Pro-PCB (500 nM) was activated with 20 nM thrombin and 20 nM TM for 60 min at 25°C. The thrombin activity was subsequently quenched with 300 nM hirudin. PCB was diluted to 10 nM in HBSP containing 2 mM CaCl 2 and 100 nM plasmin. After a 60-min incubation at 25°C, the PCB activity was quenched with 5 M Plummer's inhibitor. Ten nM plasmin Ϯ PCB treatment was incubated at room temperature with 50 nM pro-PCB in HBSP buffer containing 2 mM CaCl 2 and 5 M Plummer's inhibitor. Twenty five-l aliquots were removed at increasing times and quenched with concentrated Laemmli loading buffer (50 l) containing 2-mercaptoethanol. Twenty-l aliquots were electrophoresed on a 12% Tris glycine SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. PCB activation was monitored by Western blot analysis as described above. The developed blots were evaluated by densitometric scanning using an Arcus II Scanner (AGFA). Data were analyzed with Adobe Photoshop software and quantified with the NIH Image program (Macintosh version).
Steady-state Kinetics of Plasmin-mediated Pro-PCB Activation-Pro-PCB (15-1000 nM) was added to HBSP containing 2 mM CaCl 2 , 5-10 nM plasmin, Ϯ200 units/ml UH, and 5 M Plummer's inhibitor (50 l final volume). The reactions were performed at room temperature. Samples with and without UH were quenched at 15 and 30 min, respectively, by adding 20 l of 5-fold concentrated Laemmli loading buffer. Twenty-l aliquots were electrophoresed on 12% Tris glycine SDS-polyacrylamide gels. Samples were transferred to nitrocellulose sheets, and the 35-kDa PCB subunit was visualized by Western blot analysis. The developed blots were evaluated by densitometric scanning as described above. PCB that was generated from pro-PCB by thrombin:TM was used as the standard.
Effect of Heparin-like Polysaccharides on Plasmin-mediated Pro-PCB Activation-Samples (50-l final volume) containing 50 nM pro-PCB, 2 mM CaCl 2 , 5 nM plasmin, 0.5 mg/ml polysaccharides, and 5 M Plummer's inhibitor in HBSP buffer were placed at room temperature for 15 min. PCB was assayed by Western blot analysis and densitometry as described above.
Activation of Pro-PCB in Plasma-Biotinylated PCI (Biot-PCI) was prepared with PCI and Sulfo-NHS-LC-Biotin (Pierce) in phosphatebuffered saline. The concentration of Biot-PCI was estimated using ⑀ 280 nm ϭ 1.27 ϫ 10 4 M Ϫ1 cm Ϫ1 as deduced from the primary sequence of PCI. Pooled citrated human plasma (700 l) was mixed with 200 nM Biot-PCI (final volume ϭ 1 ml). Samples were also run in the presence of UH (1 mg/ml) or anti-␣ 2 AP monoclonal IgG (150 g/ml). Pro-PCB activation was triggered by adding t-PA (5 g/ml), and the reactions were performed at 37°C. Aliquots (200 -300 l) were removed at increasing times, quenched with D-Phe-L-Phe-L-Arg-chloromethyl ketone (200 M), transferred to 100 l of 400 nM Biot-PCI and 100 l (1:1 slurry of beads:buffer) of S-ultralink immobilized streptavidin plus beads (Pierce), and rotated at room temperature for 2 h. Samples were centrifuged, and the supernatants were assayed for Plg activation by Western blot analysis using an anti-human Plg IgG. The pellets were washed with HBS, lyophilized, resuspended in SDS sample buffer, and assayed for PCB formation by Western blot analysis using the anti-pro-PCB IgG.

Pro-PCB Is Activated by Plasmin to Yield PCB-Treatment
of 50 nM pro-PCB with 10 nM plasmin generates PCB activity as measured by hydrolysis of FA-AR (Fig. 1, plot D). In contrast, 10 nM thrombin does not appear to activate pro-PCB during the 1-h treatment period (plot A). Thrombin-mediated activation of pro-PCB is markedly increased in the presence of 10 nM TM (plot H), as shown previously (7). TM does not augment plasmin-mediated pro-PCB activation (plot E).
Pro-PCB activation was also assessed by Western blotting analysis using anti-pro-PCB IgG (Fig. 2). Pro-PCB migrates as a single band with an M r of 60,000 (lane 1). Treatment of pro-PCB with 10 nM thrombin for 10 or 30 min does not alter the mobility of the pro-PCB band (lanes 2 and 3, respectively). However, exposure of pro-PCB to thrombin:TM for 10 and 30 min (lanes 4 and 5, respectively) results in the loss of the 60-kDa band and the concomitant appearance of a 35-kDa PCB subunit protein band. Treatment of pro-PCB with plasmin for 10 and 30 min (lanes 6 and 7, respectively) generates a PCB subunit band that co-migrates with that produced by thrombin:TM.
N-terminal sequencing of the 35-kDa PCB subunits generated by thrombin:TM or plasmin was performed to map the actual cleavage sites. The first 7 amino acids obtained for the PCB subunit formed by the action of thrombin:TM (Ala-Ser-Ala-Ser-Tyr-Tyr-Glu) correspond to the sequence previously reported (4) for the thrombin-generated subunit. The plasmincatalyzed PCB subunit yielded the same N-terminal sequence; hence, both plasmin and thrombin:TM cleave pro-PCB at Arg 92 -Ala 93 .
Heparin Stimulates Plasmin-mediated Pro-PCB Activation-The presence of UH (100 units/ml) during the incubation of pro-PCB and plasmin (along with FA-AR) increases the rate of hydrolysis of the PCB substrate (Fig. 1, plot F). UH does not promote thrombin-mediated pro-PCB activation (Fig. 1, plot B). The stimulatory effect of UH on plasmin-mediated pro-PCB activation is blocked by histidine-rich glycoprotein, a heparinbinding protein, in a dose-dependent manner (data not shown). UH does not alter the activity of PCB toward FA-AR (data not shown). The insensitivity of plasmin-catalyzed small substrate (S-2390) hydrolysis to the presence of UH (data not shown) together with the previously observed binding of UH to pro-PCB (14) are consistent with the view that binding of UH to pro-PCB increases the susceptibility of pro-PCB to plasmincatalyzed activation.
Western blot analysis provides direct evidence that UH increases the rate of plasmin-mediated processing of pro-PCB. Note the more abundant 35-kDa PCB subunit at 10 and 30 min in the presence of 10 units/ml UH (Fig. 2, lanes 8 and 9, respectively) or 100 units/ml (lanes 10 and 11, respectively) when compared with the absence of UH (lanes 6 and 7, respectively). The rates of plasmin-mediated activation of pro-PCB are similar in the presence or absence of calcium, both with and without UH (data not shown). In contrast to this calciumindependent process, pro-PCB activation by thrombin:TM requires calcium (15).
Plasmin-mediated pro-PCB activation was evaluated in the presence of increasing concentrations of UH up to 200 units/ml (Fig. 3). The stimulatory effect of UH is dose-dependent and saturable. A prominent enhancement of pro-PCB activation is observed with as little as 5 units/ml of UH (plot C). Western blot analysis confirmed that the saturation kinetics does not reflect pro-PCB depletion at the higher UH concentrations (data not shown).
UH Delays Spontaneous Inactivation of PCB Activity-Incubation of PCB at 25°C leads to spontaneous loss of its catalytic activity (7,16,17). The t1 ⁄2 for this inactivation process is 74 min (Fig. 4). The t1 ⁄2 for the spontaneous inactivation of PCB is increased in the presence of UH (200 units/ml) by 2.3-fold to 170 min. This result suggests that PCB binds UH as does pro-PCB.
PCB Releases Arg from Plasmin and Reduces Its Pro-PCB Processing Activity-Incubation of plasmin with catalytic amounts of PCB at 37°C results in the release of Arg (Fig. 5). Release of 1 eq of Arg was expected since Arg is situated at the C terminus of the plasmin A subunit; however, release of greater than stoichiometric amounts of Arg is observed. Re- lease of Arg is slightly accelerated in the presence of UH. Incubation of PCB and active-site blocked plasmin, FFR-Pln, results in a markedly different Arg release profile: the amount of released Arg reaches approximately 0.7 mol of Arg per mol of FFR-Pln. These observations show that plasmin autoproteolysis (a process that should not occur with FFR-Pln) generates additional C-terminal Arg residues that are susceptible to PCB-mediated release.
Western blot analysis shows that the rate of plasmin-catalyzed formation of the 35-kDa subunit from pro-PCB is decreased by approximately 3-fold upon pretreatment of the plasmin with PCB (compare lanes 1-5 and 6 -10 in Fig. 6A, and the filled and open symbols in Fig. 6B). Our data showing that PCB catalyzes the release of Arg from plasmin and compromises the ability of plasmin to process pro-PCB corroborate recent results from Nesheim and co-workers (17).
The K m Value of Plasmin for Pro-PCB Is Similar to the Plasma Pro-PCB Concentration and the k cat Value Is Increased by UH-Densitometric scanning of the 35-kDa PCB subunit on Western blots was used to derive the kinetic constants for the activation of pro-PCB by plasmin in the absence and presence of UH. The reactions were performed in the presence of the PCB inhibitor, Plummer's inhibitor, to block the inhibitory effect of PCB activity on plasmin-mediated pro-PCB activation (as described above). The K m , k cat , and k cat /K m values of plasmin for pro-PCB are 55 nM, 4.4 ϫ 10 Ϫ4 s Ϫ1 , and 0.008 M Ϫ1 s Ϫ1 , respectively (Table I). Interestingly, the K m value of plasmin for pro-PCB is comparable to the plasma pro-PCB concentration, 50 -70 nM (1,18). The data listed in Table I also indicate that the catalytic efficiency (k cat /K m ) of plasmin toward pro-PCB is approximately 8-fold greater than that of thrombin alone. Moreover, UH produces a 16-fold increase in catalytic efficiency of plasmin-catalyzed processing of pro-PCB, due largely to an increase in the value of k cat .
Other Polysaccharides Also Promote Plasmin-mediated Pro-PCB Activation-The propensity of UH to stimulate plasminmediated pro-PCB activation is mimicked by LMWH, the UH derivative (M r ϭ 4,400) that is prepared by controlled chemical degradation (Fig. 1, plot G). LMWH does not promote thrombin-mediated pro-PCB activation (Fig. 1, plot C). A variety of other polysaccharides were also evaluated for their effects on plasmin-mediated pro-PCB activation (Fig. 7). The final concentration of each agent was 0.5 mg/ml. All values are compared with that exhibited by UH (assigned a value of 100%). The two LMWH species (LMWH 4400 and LMWH 3000) display cofactor activities that are similar to UH. Chondroitin sulfate B (i.e. dermatan sulfate) is superior to both chondroitin sulfate A and chondroitin sulfate C. Dextran sulfate 5000 and 8000 display cofactor activities that are greater than UH. Heparan sulfate and keratan sulfate do not stimulate plasminmediated pro-PCB activation at 0.5 mg/ml but do so at higher concentrations (data not shown). The apparent potency differences depicted in Fig. 7 may well reflect differences in the affinities of the polysaccharides for pro-PCB. Plasmin Activates Pro-PCB in a Plasma Milieu-PCB that is generated de novo in plasma can be detected by sequestration with Biot-PCI, capture of PCB:Biot-PCI by immobilized streptavidin, and Western blot analysis using the anti-pro-PCB IgG. The appearance of the 35-kDa band is indicative of pro-PCB activation. The addition of t-PA to Biot-PCI-spiked normal human plasma results in a progressive time-dependent accumulation of PCB (Fig. 8A, lanes 1-3). The extent of the pro-PCB activation is increased in the presence of UH (lanes 4 -6). The accumulation of the 35-kDa band is markedly enhanced when the reaction is performed in the presence of a neutralizing antibody versus ␣ 2 AP (lanes 7-9). Hence, these data reveal that de novo generated plasmin is responsible for pro-PCB activation in t-PA spiked plasma and that the reaction in plasma is accelerated by UH.
The extent of Plg activation and concomitant plasmin formation in t-PA-spiked plasma was assessed by Western blot analysis using anti-Plg IgG (Fig. 8B). Only a small fraction of Plg is converted to plasmin in t-PA-spiked plasma as judged by the formation of relatively low amounts of the plasmin-␣ 2 AP complex (Fig. 8B, lanes 2-4). The presence of UH does not substantially influence the rate of Plg activation (lanes 5-7). This analysis reveals that relatively low amounts of de novo generated plasmin are able to activate pro-PCB in a plasma milieu. The presence of the anti-␣ 2 AP IgG causes a marked decrease in the intensity of the Plg band. The absence of a corresponding increase in the plasmin-␣ 2 AP complex is expected since this antibody neutralizes ␣ 2 AP. The accelerated Plg activation observed in the presence of anti-␣ 2 AP IgG probably arises because the "longer lived" plasmin activity (due to neutralization of ␣ 2 AP) converts Glu-Plg to Lys-Plg which is more readily activated by t-PA (19).
Plasmin-mediated Activation of Pro-PCB Modulates in Vitro Clot Lysis-The predicted antifibrinolytic effect due to UHaccelerated, plasmin-mediated pro-PCB activation was explored in plasma clot lysis experiments. Plasma clot formation was triggered with batroxobin instead of thrombin because the latter is readily neutralized by antithrombin III in the presence of UH. Moreover, prothrombin-deficient plasma was used in order to assess the potential antifibrinolytic activity of UH without a confounding influence of de novo generated thrombin. The addition of batroxobin to recalcified citrated prothrombin-deficient plasma containing a t-PA spike causes an increase in turbidity that is coincident with clot formation (Fig. 9). The time to reach maximal turbidity (T init ) is approximately 30 min. In turn, plasmin that is generated by the action of t-PA on endogenous Plg lyses the plasma clot and decreases the turbidity. The time from maximal turbidity until 50% approach to base line (⌬T) is approximately 30 min.
The T init is shortened from 30 to 10 min in the presence of 25 or 100 units/ml UH. Conversely, the ⌬T values in the presence of 25 and 100 units/ml UH are prolonged from 30 to 90 and 130 min, respectively. The ability of UH to delay clot lysis is blocked by the PCB inhibitor, PCI, in a dose-dependent manner (compare the lysis profiles with 10 and 100 nM  PCI in Fig. 9). PCI in the absence of UH has a much less pronounced stimulatory effect on the rate of clot lysis. This observation may reflect enhanced susceptibility of the uncross-linked fibrin I clot (formed by batroxobin) to plasmin digestion and/or markedly reduced conversion of pro-PCB to PCB in the absence of UH. The ability of PCI to suppress the UH-induced prolongation of clot lysis using prothrombindeficient plasma shows that UH is stimulating pro-PCB activation in a thrombin-independent manner. The results described earlier indicated that de novo generated plasmin is mediating the activation of pro-PCB in t-PA-spiked plasma. Hence, our studies establish that plasmin-mediated fibrinolysis is suppressed by a negative feedback mechanism involving plasmin-catalyzed pro-PCB activation. DISCUSSION PCB suppresses fibrinolysis by releasing C-terminal basic amino acids from partially degraded fibrin thereby down-regulating plasmin production. The UH-stimulated, plasmin-catalyzed activation of pro-PCB thus exerts a negative feedback effect that can limit plasmin formation in vivo. The potential physiological importance of plasmin as a pro-PCB activator is indicated by the steady-state kinetic parameters. The apparent K m values for plasmin-catalyzed activation of pro-PCB in the absence and presence of UH are 55 and 20 nM, respectively; both of these values are similar to the plasma concentration of pro-PCB, 50 -70 nM (1,18). In contrast, the apparent K m value for thrombin-catalyzed activation of pro-PCB, 0.5-2 M, is 10-40-fold above the plasma concentration of pro-PCB (7,20). Further studies are needed to determine the relative contributions to the in vivo processing of pro-PCB by thrombin, plasmin, and their various macromolecular complexes.
We postulate that the binding of UH to pro-PCB alters the conformation and/or charge of pro-PCB thereby making it a more favorable substrate for plasmin. The purported avidity of pro-PCB for UH is concordant with the use of heparin-Sepharose for pro-PCB purification (14). It is noteworthy that UH does not facilitate thrombin-mediated pro-PCB activation. Pro-PCB contains a region (Trp 210 -Ser 221 ; WKKMRMWRKNRS, where italics indicate basic residues) that is homologous to the heparin-binding sequences, -XBBXBX-(where B and X represent basic and hydrophobic residues, respectively), in GAGbinding proteins (21). UH may promote plasmin-mediated ac-tivation of pro-PCB via an interaction with this domain in pro-PCB.
The antifibrinolytic effect of UH that stems from its ability to promote plasmin-mediated pro-PCB activation could in theory compromise the pharmacologic utility of UH for thrombotic disorders. The intended anticoagulant effect of UH due to its propensity to serve as a cofactor for antithrombin III would be opposed by the ability of UH to work in concert with plasmin to promote pro-PCB activation. The typical efficacious anticoagulant plasma level of UH (22) is somewhat less than the threshold value, 5 units/ml, that exerts an unambiguous effect on plasmin-mediated pro-PCB activation. Nevertheless, the potential impact of this clot-stabilizing effect of UH especially during aggressive anticoagulant therapy with UH is deserving of further investigation.
Structural features of polysaccharides that promote plasminmediated pro-PCB activation can be gleaned from the results shown in Fig. 7. Heparan sulfate which exhibits lower cofactor activity than heparin has fewer O-or N-linked sulfate groups. Dextran sulfate displays the greatest cofactor activity; it is highly decorated with sulfate groups. Keratan sulfate, which contains a galactose residue with a neutral 6-hydroxymethyl group, is a weaker stimulator than chondroitin sulfates A or B, which possess glucuronic acid sugar units with an anionic 6-carboxylate group. These comparisons suggest that the anionic character of the polysaccharide is an important structural determinant of the cofactor activity. The stereochemistry of the polysaccharide also appears to influence the cofactor activity. For instance, chondroitin sulfate B, which contains an iduronic acid with the 5R configuration, is a better cofactor than chondroitin sulfates A and C, which contain glucuronic acid residues with the 5S configuration.
UH is a heterogeneous mixture of heavily sulfated, long chain acidic GAGs that is isolated from a variety of different tissue sources. In particular, GAGs are synthesized by endothelial cells and deposited both in the extracellular matrix and on the abluminal cell surface. We propose that plasmin-mediated pro-PCB activation and its stimulation by GAGs plays a key role in clot stabilization at sites of vascular injury. Premature lysis of the hemostatic clot at a breach in the integrity of the vasculature must be deterred to avoid hemorrhage. Exposure of GAGs due to vascular damage may well provide a nidus for plasmin-mediated pro-PCB activation and help satisfy this need. The GAG-stimulated, plasmin-dependent mechanism for pro-PCB activation is particularly relevant at injury sites where the endothelial cell surface is denuded and TM will be absent. On the other hand, it is desirable to suppress clot extension over adjacent "uninjured" regions that contain an intact endothelial cell surface. At these sites, plasmin-mediated pro-PCB activation may be latent due to the limited GAG exposure, and the fibrin deposited in this region would have a reduced stability. GAG-mediated enhancement of plasmin-catalyzed activation of pro-PCB could thus serve to modulate the life time of fibrin clots in accordance with their proximity to the extracellular matrix thereby promoting hemostasis. FIG. 9. UH delays plasma clot lysis by promoting pro-PCB activation. Prothrombin-deficient plasma was mixed with batroxobin (5 g/ml), 9.5 mM CaCl 2 , t-PA (0.05 g/ml), and the following other additions: E, none; Ⅺ, 25 units/ml UH; q, 100 units/ml UH; ࡗ, 100 units/ml UH ϩ 10 nM PCI; OE, 100 units/ml UH ϩ 100 nM PCI; छ, 100 nM PCI. Clot formation and lysis was evident by turbidity changes that were detected at A 405 nm .