Analogs of human plasminogen that are labeled with fluorescence probes at the catalytic site of the zymogen. Preparation, characterization, and interaction with streptokinase.

Fluorescent analogs of the proteinase zymogen, plasminogen (Pg), which are specifically inactivated and labeled at the catalytic site have been prepared and characterized as probes of the mechanisms of Pg activation. The active site induced non-proteolytically in Pg by streptokinase (SK) was inactivated stoichiometrically with the thioester peptide chloromethyl ketone, Nα-[(acetylthio)acetyl]-(D-Phe)-Phe-Arg-CH2Cl; the thiol group generated subsequently on the incorporated inhibitor with NH2OH was quantitatively labeled with the fluorescence probe, 2-((4′-iodoacetamido)anilino)naphthalene-6-sulfonic acid; and the labeled Pg was separated from SK. Cleavage of labeled [Glu]Pg1 by urokinase-type plasminogen activator (uPA) was accompanied by a fluorescence enhancement (ΔFmax/Fo) of 2.0, and formation of 1% plasmin (Pm) activity. Comparison of labeled and native [Glu]Pg1 as uPA substrates showed that activation of labeled [Glu]Pg1 generated [Glu]Pm1 as the major product, while native [Glu]Pg1 was activated at a faster rate and produced [Lys]Pm1 because of concurrent proteolysis by plasmin. When a mixture of labeled and native Pg was activated, to include plasmin-feedback reactions, the zymogens were activated at equivalent rates. The lack of potential proteolytic activity of the Pg derivatives allowed their interactions with SK to be studied under equilibrium binding conditions. SK bound to labeled [Glu]Pg1 and [Lys]Pg1 with dissociation constants of 590 ± 110 and 11 ± 7 nM, and fluorescence enhancements of 3.1 ± 0.1 and 1.6 ± 0.1, respectively. Characterization of the interaction of SK with native [Glu]Pg1 by the use of labeled [Glu]Pg1 as a probe indicated a ∼6-fold higher affinity of SK for the native Pg zymogen compared to the labeled Pg analog. Saturating levels of ϵ-aminocaproic acid reduced the affinity of SK for labeled [Glu]Pg1 by ∼2-fold and lowered the fluorescence enhancement to 1.8 ± 0.1, whereas the affinity of SK for labeled [Lys]Pg1 was reduced by ∼98-fold with little effect on the enhancement. These results demonstrate that occupation of lysine binding sites modulates the affinity of SK for Pg and the changes in the environment of the catalytic site associated with SK-induced conformational activation. Together, these studies show that the labeled Pg derivatives behave as analogs of native Pg which report functionally significant changes in the environment of the catalytic site of the zymogen.

Activation of the serine proteinase zymogen, plasminogen (Pg) 1 to form plasmin is the central event in the dissolution of blood clots by the fibrinolytic system. The physiological serine proteinases, urokinase-type plasminogen activator (uPA) and tissue-type plasminogen activator (tPA), activate Pg by cleavage of the Arg 561 -Val 562 bond in the catalytic domain of the zymogen, thereby initiating degradation of fibrin by plasmin (reviewed in Henkin et al. (1991) and Ponting et al. (1992a)). The rate of Pg activation and its physiological localization are regulated by assembly of activator-Pg, enzyme-substrate complexes bound to the fibrin surface. The rate of plasmin formation is increased as a result of coordinated interactions among the activating enzymes, Pg, and fibrin, with fibrin binding mediated substantially by lysine binding sites on tPA and Pg Nesheim et al., 1990;Liu and Gurewich, 1992;Fleury et al., 1993). The rate of Pg activation is also controlled by conformational equilibria between compact and extended forms of [Glu]Pg which are shifted to the more rapidly activated, extended forms by binding of lysine analogs (Violand et al., 1978;Urano et al., 1987Urano et al., , 1988Mangel et al., 1990;Christensen and Molgaard, 1992;Marshall et al., 1994). Reactions catalyzed by plasmin play a complex role in accelerating the rate and localizing Pg activation. Plasmin converts the activating enzymes from the single-chain to the two-chain forms Collen et al., 1986;Longstaff et al., 1992); it cleaves a 77-residue peptide from the amino terminus of [Glu]Pg to produce the more reactive, [Lys]Pg form of the zymogen (Violand and Castellino, 1976;Christensen, 1977;Lucas et al., 1983;Fredenburgh and Nesheim, 1992); and it transforms fibrin into a more effective surface and cofactor of the reactions by generating new sites for productive complex assembly (Norrman et al., 1985;Bok and Mangel, 1985;Higgins and Vehar, 1987;Fleury et al., 1993).
The bacterial protein, streptokinase (SK) activates Pg by a different mechanism than that of uPA and tPA. In this mechanism, specific binding of SK to Pg induces formation of an active catalytic site in the zymogen through a conformational change (McClintock and Bell, 1971;Reddy and Markus, 1972;Schick and Castellino, 1974). Non-proteolytic activation of Pg by SK is closely coupled with enzymatic activation of Pg by Pg-SK and plasmin-SK complexes, by their cleavage of the Arg 561 -Val 562 bond (Bajaj and Castellino, 1977;Gonzalez-Gronow et al., 1978;Davidson et al., 1990). Similar to activation by the physiological proteinases, the rate of SK-initiated plasmin formation is regulated by interactions with fibrin (Strickland et al., 1982;Chibber et al., 1985;Cassels et al., 1987), Pg conformational equilibria (Chibber and Castellino, 1986), and plasmin proteolytic reactions (Strickland et al., 1982). Studies aimed at defining the molecular mechanisms of Pg activation have been complicated by the concurrent plasmin-catalyzed reactions. This problem has been addressed recently by the use of engineered variants of Pg, uPA, and tPA, in which plasmin cleavage sites have been mutated to prevent the reactions, or the catalytic site serine residue of Pg has been replaced to render the plasmin formed inactive (Boose et al., 1989;Lijnen et al., 1990;Davidson et al., 1990;Liu and Gurewich, 1992;Fleury et al., 1993).
The present studies were undertaken to develop a new approach for investigating fibrinolytic reaction mechanisms, based on derivatives of Pg in which the catalytic site of the zymogen has been inactivated and labeled with an extrinsic fluorescence probe. To prepare these analogs, the active site generated non-proteolytically in Pg by SK binding is irreversibly inactivated by site-directed alkylation with a thioester peptide chloromethyl ketone (Bock, 1988(Bock, , 1992a. The thiol group generated subsequently at the amino terminus of the covalently incorporated inhibitor provides a unique site for selective labeling with a variety of fluorescence probes (Bock, 1992b). Dissociation of the labeled Pg-SK complex allows purification of the fluorescent derivatives of Pg. The first Pg derivatives of this type, labeled with the fluorescence probe, 2-anilinonaphthalene-6-sulfonic acid, were prepared and characterized in the present studies. The labeled [Glu]Pg derivative was found to have properties similar to native Pg and to report changes in the catalytic site of the zymogen through changes in the probe fluorescence accompanying proteolytic activation by uPA and non-proteolytic activation by SK. The absence of potential proteolytic activity of the Pg analogs allowed SK-Pg interactions to be studied under equilibrium binding conditions for the first time. The approach developed in these studies is expected to enable further definition of binding interactions and elementary reaction steps of Pg activation which are coupled to conformational changes affecting the catalytic site of the zymogen.

EXPERIMENTAL PROCEDURES
Protein Purification and Characterization- [Glu]Pg was purified from human plasma and separated into the carbohydrate variants, Pg 1 and Pg 2 , by minor modifications of published procedures (Deutsch and Mertz, 1970;Castellino and Powell, 1981).
[Lys]Pg 1 was prepared by cleavage of 87 M [Glu]Pg 1 with 7 M plasmin in 50 mM Tris-Cl, 20 mM lysine, 0.1 M NaCl, pH 9.0, for 30 min at 25°C. Plasmin was removed by adsorption onto soybean trypsin inhibitor-agarose (5 mg of inhibitor/ml of gel). Plasmin was prepared by activation of 10 M [Glu]Pg with 0.1 M uPA in 10 mM Mes, 10 mM Hepes, 0.15 M NaCl, 20 mM ⑀-ACA, 1 mg/ml PEG, pH 7.4, at 25°C. After completion of activation, as moni-tored by the increase in chromogenic substrate activity, the plasmin was bound to 25 ml of soybean trypsin inhibitor-agarose and the gel was washed in a column at 22°C with buffer of the above composition but with pH 6.0 and containing 1 M NaCl. The gel was then washed with the same pH 6.0 buffer containing 0.15 M NaCl and the plasmin was eluted with 10 mM Mes, 10 mM Hepes, 0.5 M benzamidine, 20 mM ⑀-ACA, 1 mg/ml PEG, pH 5.0. The plasmin was concentrated by ultrafiltration, dialyzed against the above Tris/lysine buffer, and stored at Ϫ70°C. For preparation of active site-blocked plasmin-Sepharose, 50 M [Glu]Pg was activated with 0.5 M SK in 1 M Hepes, 0.3 M NaCl, 1 mM EDTA, pH 7.0, at 25°C. After completion of the reaction, 82 M ATA-FFR-CH 2 Cl was added and the enzyme was inactivated by incubation for 1 h. The plasmin was dialyzed against the pH 7 buffer and coupled to SulfoLink gel (Pierce) at 7.5 mg/ml gel by incubation for 90 min in the presence of 0.1 M NH 2 OH. The gel, containing ϳ7 mg of plasmin/ml, was washed with buffer containing 3 M NaSCN before use. Pure recombinant twochain uPA was a gift from Dr. Jack Henkin (Abbott Laboratories). SK was obtained from Kabi (Kabikinase) as the pure protein or was purified from outdated therapeutic preparations by chromatography on plasmin-Sepharose. Homogeneous SK was obtained by chromatography on the gel in 0.1 M Hepes, 0.1 M NaCl, 1 mM EDTA, 1 mg/ml PEG, pH 7.4, elution with this buffer containing 3 M NaSCN, and dialysis exhaustively against buffer lacking NaSCN. Protein concentrations were determined from the 280-nm absorbance using the following extinction coefficients ((mg/ml) Ϫ1 cm Ϫ1 ) and molecular weights: [Glu]Pg, 1.69, 92,000; [Lys]Pg and plasmin, 1.70, 84,000 (Sjoholm et al., 1973;Violand and Castellino, 1976;Castellino and Powell, 1981); SK, 0.95, 47,000 (Taylor and Botts, 1968;Jackson and Tang, 1982). Concentrations of uPA and plasmin were determined by active-site titration with fluorescein-mono-p-guanidinobenzoate (Melhado et al., 1982;Bock et al., 1989).
Gel Electrophoresis-SDS-polyacrylamide gels of constant polyacrylamide percentage or a 4 -15% gradient (Bio-Rad) were run in the Laemmli buffer system (Laemmli, 1970) and stained with Coomassie Blue R-250 or with colloidal Coomassie Blue (Novex). Labeled proteins were visualized with a 300-nm transilluminator. Samples were prepared as described previously (Bock, 1993) except that aliquots of proteins in 1 M Hepes buffer were diluted 10-fold with water before sample preparation to prevent separation anomalies.
Chromogenic Substrate Assays-Plasmin activity was measured from the initial rate of hydrolysis of 200 M H-D-Val-Leu-Lys-p-nitroanilide (S2251, KabiVitrum) monitored by the linear absorbance increase at 405 nm with time in 0.1 M Hepes, 0.1 M NaCl, 1 mM EDTA, 1 mg/ml PEG, Ϯ10 mM ⑀-ACA, pH 7.4, at 25°C. Kinetic parameters for active site-titrated plasmin with S2251 of k cat ϭ 18.7 Ϯ 1.6 s Ϫ1 and K m ϭ 110 Ϯ 14 M in the absence of ⑀-ACA were obtained by fitting the Michaelis-Menten equation to initial rates at 20 -1500 M S2251.
Preparation of Labeled Pg-ATA-FFR-CH 2 Cl was prepared as described previously (Bock, 1992a(Bock, , 1993. Inactivation of the Pg-SK complex was typically performed by adding 375-400 M ATA-FFR-CH 2 Cl to 25 M Pg in 1 M Hepes buffer, 0.3 M NaCl, 1 mM EDTA, 1 mg/ml PEG, pH 7.0, and initiation of the reaction at 25°C by addition of 50 M SK. To ensure completion of the reaction, as measured by the loss of chromogenic substrate activity (Ͻ0.1% active), incubation was continued for 1.5-3 h. The reaction mixture was chromatographed on Sephadex G-25 (2.5 ϫ 50 cm) at 4°C to remove the excess inhibitor. Labeling with fluorescence probes was carried out by addition of 110 -160 M 5-IAF or IAANS (Molecular Probes) to 6 -26 M Pg-SK complex, representing a 5-13-fold molar excess of probe, in the above 1 M Hepes, pH 7, buffer and initiation of the reaction by addition of 0.1 M NH 2 OH, following procedures described previously (Bock, 1992a(Bock, , 1992b(Bock, , 1993. After incubation for 1 h in the dark at 25°C, the excess dye was removed by chromatography on Sephadex G-25. The labeled Pg-SK complex was incubated overnight at 25°C with 50 M FFR-CH 2 Cl to inactivate any residual enzyme activity, chromatographed on plasmin-Sepharose (2.5 ϫ 8 -12 cm) in the 1 M Hepes, pH 7, buffer to remove excess SK, and dialyzed into 0.1 M Hepes, 0.1 M NaCl, 1 mM EDTA, 1 mg/ml PEG, pH 7.4. The sample was concentrated by adsorption onto a small column (1.5 ϫ 7 cm) of lysine-Sepharose using batchwise elution with buffer containing 0.2 M ⑀-ACA. The Pg-SK complex was dissociated by dialysis against the 0.1 M Hepes, pH 7.4, buffer containing 3 M NaSCN and the proteins were separated by gel filtration in this buffer at 4°C on a column (2.5 ϫ 112 cm) of Sephacryl S200 HR. Fractions (1-3 ml) containing labeled Pg, partially resolved from the overlapping peak of SK, were made 50 M FFR-CH 2 Cl, pooled, dialyzed, and concentrated, and rechromatographed on Sephacryl S200 HR. A second method was also used for isolation of labeled Pg from the Pg-SK complex. The Pg-SK complex was dialyzed against 0.1 M Hepes, 0.1 M NaCl, 10 mM ⑀-ACA, 1 mM EDTA, 0.02% azide, pH 7.4, and chromatographed at room temperature on plasmin-Sepharose (1.5 ϫ 16 cm) at a flow rate of 2 ml/h. Under these conditions, labeled Pg eluted with the equilibration buffer and SK partitioned onto the affinity matrix. The separation was repeated after elution of SK from the gel with buffer containing 3 M NaSCN. Residual labeled plasmin and [Lys]Pg were removed by aminohexyl-Sepharose (Pharmacia or Sigma) chromatography, following published procedures (Nieuwenhuizen and Traas, 1989). The purified, labeled Pg was dialyzed into the above 0.1 M Hepes, pH 7.4, buffer and stored at Ϫ70°C.
[AANS]FFR-[Lys]Pg was prepared similarly except that after the S200 HR chromatography step described above, the labeled Pg was chromatographed at 22°C on a 15-ml column of SK-Affi-Gel 10 to remove residual plasmin and SK. SK was coupled to Affi-Gel 10 essentially as recommended by Bio-Rad, with ϳ5 mg of SK/ml of gel.
[AANS]FFR-[Lys]Pg bound to the gel in 0.1 M Hepes, 0.1 M NaCl, 10 mM ⑀-ACA, 1 mM EDTA, 1 mg/ml PEG, pH 7.4, and was eluted as a broad peak with buffer containing 1.5 M NaCl, separated from SK-plasmin complex, which did not bind to the gel, and free plasmin, which remained bound.
Quantitation of Inhibitor and Probe Incorporation-Incorporation of ATA-FFR-CH 2 Cl into the Pg-SK complex was quantitated with 5,5Јdithiobis(2-nitrobenzoic acid) following previously described procedures (Bock, 1988(Bock, , 1993. Reactions were performed in 1 M Hepes, 0.3 M NaCl, 1 mM EDTA, 1 mg/ml PEG, pH 7.0, with ATA-FFR-(Pg-SK) complex that had been freed of excess inhibitor. The free thiol concentration resulting from spontaneous thioester hydrolysis was calculated from the difference in 412 nm absorbance between a sample and a blank lacking the protein after addition of 500 M 5,5Ј-dithiobis(2-nitrobenzoic acid), using an extinction coefficient of 14,150 M Ϫ1 cm Ϫ1 (Riddles et al., 1979). The thioester concentration was then obtained from the amplitude of the thiol burst reaction initiated by addition of 0.09 M NH 2 OH. The stoichiometry of inhibitor incorporation was calculated from these measurements and the 280-nm absorbance of the Pg/SK mixture, assuming additivity of the protein extinction coefficients. Incorporation of IAANS was quantitated from the probe and protein absorbances in 0.1 M Tris-Cl, 6.0 M guanidine, 1 mM EDTA, pH 8.5, with an extinction coefficient of 26,600 M Ϫ1 cm Ϫ1 at 328 nm and ⑀ 280 nm /⑀ 328 nm ϭ 0.78 (Bock, 1992b). The Pg concentration was calculated from the corrected 280 nm absorbance with an extinction coefficient of 1.6 ((mg/ ml) Ϫ1 cm Ϫ1 obtained in the Tris/guanidine buffer.
Fluorescence Studies-Fluorescence measurements were made with an SLM 8000 spectrofluorometer in the ratio mode. Proteins were in 0.1 M Hepes, 0.1 M NaCl, 1 mM EDTA, 1 mg/ml PEG, pH 7.4, Ϯ10 M FFR-CH 2 Cl and Ϯ⑀-ACA where indicated. Measurements were made at 25°C in acrylic cuvettes (Sarstedt) coated with polyethylene glycol 20,000 to prevent protein adsorption (Latallo and Hall, 1986). Emission spectra, titrations, and kinetics experiments were performed with excitation at 340 nm (4-or 8-nm band pass). For titrations, the fluorescence change at the approximate emission difference maximum of 449 nm was measured (4-or 8-nm band pass) following successive additions of small volumes of SK, with Ͻ10% dilution. Measurements were corrected for background and scattering by subtraction of readings from a blank to which SK was added in parallel. The relatively low fluorescence yield of free [AANS]FFR-Pg resulted in blank measurements that typically represented 30 -35% of the initial fluorescence at 0.2 M labeled Pg. Fluorescence data were expressed as the fractional change in the initial fluorescence Titrations as a function of the total concentration of SK ([SK] T ) were fit by the quadratic equilibrium binding equation, where [AANS]FFR-Pg is represented by Pg, to obtain the maximum fluorescence change (⌬F max /F o ) and the dissociation constant (K D ). For this analysis, SK was assumed to bind to a single site on Pg (n ϭ 1).
Least-squares fitting was performed with the computer program SCI-ENTIST (MicroMath Software). All reported estimates of error represent Ϯ 2 S.E. The kinetics of the fluorescence changes following addition of uPA to labeled Pg were measured under the same conditions described above. Measurements were corrected for background by use of a blank lacking the labeled protein, and for long reactions, corrected for changes (Յ10%) in fluorescence of an identical reference mixture containing labeled Pg and 10 M FFR-CH 2 Cl that was incubated in parallel. The effect of native Pg on SK binding to labeled Pg was measured from the fluorescence changes recorded with time following addition of SK to preequilibrated mixtures of [Glu]Pg 1 and [AANS]FFR-[Glu]Pg 1 . Two traces from identical reactions were averaged, and the amplitudes of the rapidly established, initial fluorescence changes were obtained from the blank-corrected data collected within the first 18 -50 s of the lag phase. The fluorescence amplitudes as a function of native Pg concentration were analyzed as competitive binding of labeled and native Pg to SK to obtain the dissociation constant for SK binding to [Glu]Pg 1 by least-squares fitting of the cubic binding equation described previously Lindahl et al., 1991). For this analysis, the dissociation constant (590 Ϯ 100 nM) and maximum fluorescence change (2.8 Ϯ 0.1) for SK binding to labeled Pg were fixed at the values independently determined for the same protein preparations.

Specificity of Pg
Labeling-To label the catalytic site in Pg, the active site generated by SK was inactivated with the thioester peptide chloromethyl ketone, ATA-FFR-CH 2 Cl, and the thiol group generated subsequently by reaction of ATA-FFR-(Pg-SK) with NH 2 OH was labeled with a fluorescence probe iodoacetamide. To carry out these studies it was necessary to overcome the problem of the lower solubility of the Pg-SK complex, relative to that of the separate proteins. High concentrations of Hepes increased the solubility of the complex from typically Յ5 M in neutral pH buffers to Ն25 M in 1 M Hepes, 0.3 M NaCl, 1 mM EDTA, pH 7.0, at 25°C. Equivalent concentrations of NaCl did not produce the same effect. 2 The specificity of [Glu]Pg labeling in 1 M Hepes buffer at pH 7.0 and 25°C was examined by quantitating inhibitor and probe incorporation, and by the appearance of probe fluorescence associated with Pg on SDS-polyacrylamide gels. Addition of 50 M SK and a 15-fold molar excess of ATA-FFR-CH 2 Cl to 25 M [Glu]Pg resulted in complete inactivation (Ͻ0.1% active) of the chromogenic substrate activity and incorporation of 0.98 Ϯ 0.20 mol of inhibitor thioester groups/mol of Pg, as determined from results for several preparations. Although Hepes is only weakly nucleophilic, the requirement for high concentrations to achieve the needed solubility of the complex was associated with lower stability of the thioester group than has been typically observed (Bock, 1992b). Preparations of ATA-FFR-(Pg-SK) contained 4 Ϯ 5% free thiol when precautions were taken to minimize prolonged incubation in 1 M Hepes buffer. No significant differences were observed in the stoichiometry of inhibitor incorporation or the level of free thiol for preparations of inactivated [Glu]Pg-SK complex containing either a mixture of Pg forms 1 and 2 (1.01 Ϯ 0.17 mol of inhibitor/mol of Pg), or isolated form 1 (0.94 Ϯ 0.09 mol of inhibitor/mol of Pg).
In the second step of labeling, generation of the inhibitor thiol group in ATA-FFR-(Pg-SK) with NH 2 OH in the presence of 5-IAF resulted in covalent incorporation of the probe selectively into the zymogen, as shown by the SDS gel results in Fig.  1. SK or Pg treated separately with 5-IAF were not significantly labeled (Fig. 1). An additional experiment (not shown) in which Pg alone was subjected to incubations with the inhibitor and probe demonstrated that the presence of SK was required for labeling. Labeling was also prevented in other control reactions in which NH 2 OH was omitted, or when the catalytic site was blocked with FFR-CH 2 Cl prior to incubation with ATA-FFR-CH 2 Cl and 5-IAF in the presence of NH 2 OH (Fig. 1). These results indicated that ATA-FFR-CH 2 Cl was covalently incorporated specifically into the active catalytic site produced in Pg by SK binding, and that the fluorescence probe was subsequently incorporated by selective modification of the inhibitor thiol generated with NH 2 OH.
Isolation and Characterization of Catalytic Site-labeled Pg-Labeled Pg was isolated by separating the proteins of the dissociated Pg-SK complex by gel filtration in 3 M NaSCN, or by repeated affinity chromatography on active-site-blocked plasmin-Sepharose, under conditions where SK bound preferentially to the matrix (see "Experimental Procedures"). SDS-gel electrophoresis of the isolated, labeled Pg prepared by either method typically showed that the protein was primarily in the zymogen form but contained low levels of labeled plasmin (Fig.  1). This plasmin was generated during the initial step of inactivation of the Pg-SK complex and was minimized by the use of high concentrations of inhibitor. Results of densitometric analysis of Coomassie Blue-stained SDS gels of labeled Pg preparations showed that contamination by SK was Ͻ1%. As shown in Fig. 1 electrophoresis, the fluorescence enhancement, and the appearance of plasmin chromogenic substrate activity (Fig. 3). The labeled and native Pg reactions were accompanied by the appearance of initial cleavage products corresponding to the heavy and light chains of [Glu]Pm 1 and [Lys]Pm 1 , respectively (Fig. 3, A and B, insets). The maximum chromogenic substrate activity developed with cleavage of [AANS]FFR-[Glu]Pg 1 was 1% of that generated with native Pg (Fig. 3A). This was consistent with formation of [Glu]Pm 1 as the major product of labeled Pg cleavage because of the low level of active plasmin produced, and generation of [Lys]Pm 1 as the product of native Pg activation and associated proteolysis by plasmin. Prolonged reaction of the labeled Pg was accompanied by slow appearance of [Lys]Pm 1 due to the small amount of plasmin generated. Formation of [Lys]Pm 1 from native Pg occurred ϳ5-fold more rapidly than [Glu]Pm 1 generation from labeled Pg (Fig. 3, A  and B). To determine whether production of active plasmin accounted for the difference in the rates observed, activation of native and labeled Pg were compared in a mixture of the proteins, with the labeled protein representing a small fraction (9%) of the total Pg. The reaction was done in the presence of 10 mM ⑀-ACA, where [Glu]Pm 1 did not accumulate appreciably as an intermediate, and where the rate of native Pg activation was enhanced ϳ2-fold. Under these conditions, the rate of activation of labeled Pg was enhanced ϳ10-fold and the time courses of the activity and fluorescence changes, representing conversion of native and labeled Pg to the corresponding [Lys]Pm 1 forms, were superimposable (Fig. 3C). This indicated that [AANS]FFR-[Glu]Pg 1 had properties similar to native [Glu]Pg 1 as a substrate of uPA in the overall activation process, i.e. in the presence of accompanying plasmin feedback reactions.

Binding of SK to [AANS]FFR-[Glu]Pg 1 and [AANS]FFR-[Lys]Pg 1 -Addition of SK to [AANS]FFR-[Glu]
Pg 1 resulted in a large increase in the probe fluorescence and a blue shift of ϳ12 nm in the emission spectrum, similar to that seen for uPA activation ( Fig. 2A). The magnitude of the fluorescence enhancement at 449 nm with near-saturating levels of SK was 35% larger than that observed for activation by uPA.  (1.3-3.3). This was traced primarily to differences in the levels of labeled plasmin in early preparations and the effect of its more intense fluorescence relative to Pg (see Fig. 2A). 3 These effects were reduced but not entirely eliminated for preparations of labeled [Glu]Pg 1 in which plas- min was removed and which appeared pure by SDS-gel electrophoresis (see Fig. 2). The source of the residual variation between preparations in the amplitudes of the fluorescence enhancements, which ranged from 2.4 to 3.1, independent of the dissociation constants (280 -630 nM), is thought to be remaining differences in the levels of minor labeled species and possibly differences in the microheterogeneity of the initial Pg preparations (Deutsch and Mertz, 1970;Wallen and Wiman, 1970;Pirie-Shepherd et al., 1995). The results obtained with single preparations of labeled [Glu]Pg 1 and [Lys]Pg 1 are shown here to allow direct comparison of the amplitudes of the fluorescence changes.
The possible role of lysine binding sites in the interaction of SK with Pg was examined from the effect of ⑀-ACA on SK binding. Analysis of fluorescence titrations of [AANS]FFR-[Glu]Pg 1 in the presence of 10 or 100 mM ⑀-ACA showed a decrease in the amplitude of the fluorescence change from a maximum of 3.1 Ϯ 0.1 in the absence of ⑀-ACA to 1.6 Ϯ 0.1 at 10 mM and 1.8 Ϯ 0.1 at 100 mM ⑀-ACA. Corresponding dissociation constants of 820 Ϯ 160 and 1160 Ϯ 150 nM were obtained, showing saturation of the effect with a modest, ϳ2-fold decrease in affinity (Fig. 4).
Results  (Fig. 5). The dissociation constant for SK binding to labeled [Lys]Pg 1 was 11 Ϯ 7 nM, showing a ϳ54-fold higher affinity than for [Glu]Pg 1 . Because of the higher affinity, analysis of this data also allowed estimation of a stoichiometry for this interaction of 1.0 Ϯ 0.2 mol of SK/mol of Pg. In further contrast to the results for [Glu]Pg 1 , ⑀-ACA greatly decreased the affinity of SK binding to [AANS]FFR-[Lys]Pg 1 , while having little effect on the amplitude of the fluorescence change. As shown by analysis of the data in Fig. 5, the presence of 10 or 100 mM ⑀-ACA resulted in SK binding to labeled [Lys]Pg 1 with dissociation constants of 960 Ϯ 140 and 1080 Ϯ 100 nM, respectively, and maximum fluorescence changes of 1.9 Ϯ 0.1. These results showed saturation of the effect of ⑀-ACA with an overall ϳ98-fold decrease in affinity of SK for [Lys]Pg 1 , yielding a dissociation constant indistinguishable from that for [Glu]Pg 1 .

Results obtained with other preparations of labeled [Glu]Pg and [Lys]
Pg that were mixtures of forms 1 and 2, or which contained significant levels of labeled plasmin showed similarly large differences in affinity for SK and the differential effect of ⑀-ACA on the interactions.

Characterization of the Interaction of Native [Glu]Pg 1 with SK by the Use of [AANS]FFR-[Glu]Pg 1 as a Probe-The effect of [Glu]Pg 1 on SK binding to [AANS]FFR-[Glu]
Pg 1 was examined to further evaluate the labeled zymogen as a Pg analog and to characterize the interaction of SK with native Pg. Addition of SK to mixtures of labeled Pg and unlabeled Pg was accompanied by rapid increases in fluorescence to levels which were relatively constant during lag phases that were followed by slower time-dependent increases (Fig. 6). These kinetics allowed the amplitudes of the initial fluorescence increases to be determined, which decreased as the concentration of native Pg was increased at fixed levels of SK and labeled Pg. These results were taken to represent rapid equilibration of SK interactions with native and labeled Pg, followed by proteolytic Pg activation. The dependence of the kinetically-resolved amplitudes of the dead-time fluorescence changes on native Pg concentration was analyzed as competitive binding of labeled and native Pg to SK (Fig. 6). This analysis gave a satisfactory fit to the data with an estimate of 90 Ϯ 60 nM for the dissociation constant, indicating a ϳ6-fold higher affinity of SK for native Pg compared to the fluorescent Pg derivative.

DISCUSSION
These studies were undertaken to develop fluorescent derivatives of the plasminogen zymogen of a new type, to be used as probes for investigation of fibrinolysis mechanisms. The feasibility of the labeling strategy for preparing a family of probelabeled derivatives of the zymogen was demonstrated in the preparation and characterization of the [AANS]FFR-Pg derivatives. These derivatives were found to have properties functionally analogous to native Pg, and to report conformational changes associated with activation of the catalytic site resulting from specific proteolytic cleavage by uPA, or induced nonproteolytically by binding of SK. Irreversible inactivation of the catalytic site in the labeled Pg analogs allowed activation by uPA and equilibrium binding of SK to be studied without significant complications from plasmin-catalyzed feedback reactions. Evaluation of the specificity of labeling of the Pg-SK complex and characterization of the isolated Pg derivatives supports the conclusion that the Pg analogs were specifically labeled at the active catalytic site induced in Pg by SK binding. The results indicated a high degree of specificity of the labeling reactions, similar to that seen with other proteinases by the same method (Bock, 1992a(Bock, , 1992b. These results and the mechanism of peptide chloromethyl ketone inhibition support the idea that the inhibitor is attached to the imidazole group of the catalytic site histidine residue and the probe is attached to the aminoterminal thiol of the tripeptide inhibitor (Powers and Harper, 1986;Bock, 1992a). The isolation and characterization of the affinity-labeled active species generated by SK provided direct evidence that the active catalytic site in the Pg-SK complex is the same as the proteinase active site of plasmin, in agreement with the conclusions of previous studies (Schick and Castellino, 1974). Results of labeling experiments with the carbohydrate variants of [Glu]Pg and [Lys]Pg indicate that all of these forms can be similarly labeled with comparable specificity.
[AANS]FFR-[Glu]Pg 1 was compared to native [Glu]Pg 1 as a uPA substrate to assess its properties as an analog of the native zymogen. To our knowledge, no catalytic site-labeled serine proteinase zymogen derivatives have been previously described. Because these derivatives contain the tripeptide affinity label covalently bound at the catalytic site, it was of interest to determine whether they exhibited properties analogous to the native zymogen or to the activated proteinase.  (Violand and Castellino, 1976). Cleavage of [AANS]FFR-[Glu]Pg 1 by uPA produced [AANS]FFR-[Glu]Pm 1 as the major initial product, correlated with a large enhancement in the probe fluorescence that signaled activation of the catalytic site. Comparison of the rates of the uPA activation reactions for native and labeled Pg showed that cleavage of [AANS]FFR-[Glu]Pg 1 was slower than generation of [Lys]Pm 1 from native Pg. These results were consistent with the established role of plasmincatalyzed feedback reactions in the process (Violand and Castellino, 1976;Christensen, 1977;Violand et al., 1978;Lucas et al., 1983) and the low level (1%) of active plasmin formed from the labeled zymogen. The absence of active plasmin formation, however, prevented a direct assessment of the properties of the labeled Pg as a uPA substrate analog. To address this question, uPA activation of native and labeled Pg were compared in a mixture of the zymogens and in the presence of ⑀-ACA. Under these conditions, the rates of activation of [AANS]FFR-[Glu]Pg 1 and native [Glu]Pg 1 and the resulting time courses of their conversion to the [Lys]Pm 1 forms were indistinguishable. The observation that labeled Pg and the native zymogen behave similarly in the overall activation process, under conditions where the rate is dependent on interactions of ⑀-ACA and plasmin-catalyzed reactions, in addition to cleavage by uPA, indicates that the labeled derivatives can be employed as reporting substrate analogs of plasminogen activation.
The fluorescent Pg analogs allowed quantitative equilibrium binding studies of the SK-Pg interaction to be done for the first time. The mechanism of Pg activation by SK involves SKinduced conformational activation of the Pg catalytic site, closely coupled with irreversible proteolytic conversion of Pg to plasmin (McClintock and Bell, 1971;Reddy and Markus, 1972;Schick and Castellino, 1974;Bajaj and Castellino;1977;Gonzalez-Gronow et al., 1978;Davidson et al., 1990). Inactivation of the catalytic site in the Pg analogs allowed the binding interactions to be studied without the complications caused by subsequent proteolysis. The results of these studies provided two types of information: the affinities of binding and the amplitudes of the fluorescence changes reporting perturbations of the catalytic site-bound probe. Comparison of SK binding to labeled [Glu]Pg 1 and [Lys]Pg 1 revealed a ϳ54-fold higher affinity for [Lys]Pg 1 . Although the affinities of these interactions have not been previously quantitated, the dissociation constants of 590 Ϯ 110 nM obtained for [Glu]Pg 1 and 11 Ϯ 7 nM for [Lys]Pg 1 represented lower affinities than were anticipated on the basis of previous studies. High affinity binding was previously inferred mainly from: (i) active-site titrations of Pg activation under experimental conditions where estimation of affinity below the micromolar range would have been uncertain (McClintock and Bell, 1971;Reddy and Markus, 1972), and (ii) indirectly from the kinetic behavior of mixtures of nanomolar concentrations of SK and Pg (Wohl et al., 1980(Wohl et al., , 1983Chibber and Castellino, 1986;Davidson et al., 1990). It should be noted, however, that the affinity determined for labeled [Lys]Pg 1 in the present studies may be underestimated because the relatively low fluorescence yield of the probe limited the titrations to a concentration range in excess of the dissociation constant. The properties of the labeled derivatives as Pg analogs and reporters of SK binding were evaluated further by quantitatively characterizing the interaction between native Pg and SK using [AANS]FFR-[Glu]Pg 1 as a binding probe. The resulting estimate of 90 Ϯ 60 nM for the dissociation constant for SK binding to native [Glu]Pg 1 indicated that the presence of the label in the catalytic site resulted in a ϳ6-fold reduced affinity of SK for the Pg derivative. Such an effect is not surprising because it is thought to reflect the influence of occupation of the catalytic site by the probe-tripeptide label on the thermodynamically linked binding and conformational change which result in Pg activation and accompanying formation of the substrate binding subsites. On this basis, the affinity of SK may also be expected to vary for other active-site-liganded Pg species, depending on the structures of the bound fluorescent labels, active-site-titrants, or substrates employed to report the interaction. In this respect, it is interesting that the presence of the label decreased rather than increased SK affinity and thus did not apparently stabilize a more active enzyme-like conformation of the labeled zymogen. These observations support the conclusion that the active-site labeled Pg derivatives should be considered analogs of the native zymogen, which can be expected to exhibit functionally similar but not quantitatively identical properties. The magnitude of the effect of the presence of the label on the properties of the derivative studied here indicates that this will not compromise the use of the fluorescent Pg analogs as probes of Pg interactions and activation mechanisms.
Examination of the influence of ⑀-ACA on the affinity of SK for the fluorescent Pg derivatives revealed additional differences between [Glu]Pg 1 and [Lys]Pg 1 . The affinity of SK for [Glu]Pg 1 was reduced only ϳ2-fold by saturating levels of ⑀-ACA, while the higher affinity for [Lys]Pg 1 was reduced ϳ98fold. This differential effect resulted in indistinguishable affinities of the Pg species for SK at saturating concentrations of ⑀-ACA. These effects are presumably due to interactions of ⑀-ACA with lysine binding sites on Pg, because there is presently no evidence for such sites on SK. It has been previously shown that [Glu]Pg is in a compact conformation, which is shifted to an extended form by ⑀-ACA binding (Violand et al., 1978;Mangel et al., 1990;Ponting et al., 1992b;Christensen and Molgaard, 1992;Marshall et al., 1994). [Lys]Pg is in a more extended conformation than that of [Glu]Pg in the absence of ⑀-ACA, which is shifted to a fully extended form similar to that of [Glu]Pg in the presence of ⑀-ACA (Ramakrishnan et al., 1991;Marshall et al., 1994). These and other studies have demonstrated differences between ⑀-ACA binding to [Glu]Pg and [Lys] Pg (Markus et al., 1978;Christensen, 1984). It has been proposed that the compact conformation of [Glu]Pg is stabilized by an intramolecular interaction between the amino-terminal 77residue sequence and lysine binding sites in kringles 4 and 5 (Wiman and Wallen, 1975;Christensen, 1984;Ramakrishnan et al., 1991;Christensen and Molgaard, 1992;Marshall et al., 1994). This interaction is lost when the peptide is removed in [Lys]Pg, exposing these lysine binding sites. Thus, the high affinity of SK for [Lys]Pg and its dependence on ⑀-ACA could reflect preferential binding of SK to the partially extended conformation of [Lys]Pg and/or interaction of SK with the lysine binding sites exposed in kringles 4 and/or 5. In the latter case, the lower affinity of SK for [Glu]Pg may be due to competition with the amino-terminal peptide for these sites. Saturation of these sites by ⑀-ACA would be expected to lower selectively the affinity of SK for [Lys]Pg to that of [Glu]Pg, as is observed. Additional studies will be required to determine whether this is the basis for the differences in SK affinity.
The large enhancements in the fluorescence of the catalytic site-bound probe in [AANS]FFR-Pg that accompanied SK binding are thought to report conformational changes associated with SK-induced activation of the catalytic site and formation of the substrate-binding specificity subsites. The changes in the fluorescence spectral properties of the active-site-bound probe were similar but not the same as those accompanying proteolytic activation by uPA. Fluorescence enhancements of 0.5-1.0 have been observed for SK binding to labeled plasmin, suggesting that the difference in the amplitudes of the fluorescence changes may be accounted for by an additional perturbation of the environment of the activated catalytic site due to SK binding. The larger difference observed in the magnitude of the fluorescence changes accompanying SK binding to [Glu]Pg 1 and [Lys]Pg 1 suggests that there are also significant differences in the environments of the catalytic sites in these complexes. These differences were modulated by ⑀-ACA, as shown by the reduction of the amplitude of the fluorescence change for [Glu]Pg 1 while that for [Lys]Pg 1 was not similarly affected. Thus, although further studies will be needed to determine precisely the sources of these effects, it appears that occupation of lysine-binding sites by ⑀-ACA, Pg conformational equilibria, as well as direct effects of SK binding are linked to perturbations in the environment of the Pg catalytic site in complexes with SK that may be expected to have functional significance.
In summary, derivatives of Pg that are specifically labeled with fluorescence probes attached to the catalytic site via a tripeptide chloromethyl ketone have properties analogous to those of native Pg. The fluorescent Pg analogs provide unique opportunities to observe events at the catalytic site of the zymogen that accompany its activation. Results obtained with the [AANS]FFR-Pg analogs demonstrate the utility of the derivatives for investigation of the mechanisms of conformational and proteolytic Pg activation, under conditions where principal activation reaction steps can be separated from concurrent plasmin-catalyzed reactions. Although the present studies have focused on the Pg derivatives prepared with one fluorescence probe, the capabilities of the labeling approach indicate that this can be extended to include a family of Pg derivatives labeled with different fluorescence probes, as well as other types of labels.