Domain Truncation Studies Reveal that the Streptokinase-Plasmin Activator Complex Utilizes Long-range Protein-Protein Interactions with Macromolecular Substrate to Maximize Catalytic Turn-Over

describing interactions between soluble ligates (HPG and (cid:181)PG) and nSK/SK-domains complexed with immobilized HPG. In binary interaction studies, it was evident that when soluble nSK/SK-domain was added to the immobilized HPG, a rapid and avid nSK/SK-domains.HPG (except single domains, α and β ) binary complex formation occurs. The dissociation of these binary complexes is very slow due to the high stability of SK.HPG complex (23). After allowing the complex to dissociate maximally (~20 min), the dissociation baseline becomes stable, which remains unaffected even after washing with 2.5mM EACA. In case of single domains ( α and β ), weak binary complexation with immobilized HPG was observed, and it was seen that the dissociation rates in binding buffer itself were very high; thus, stable baselines as observed in case of bi-domains and nSK were not observed.


SUMMARY
In order to explore the inter-domain co-operativity during human plasminogen (HPG) activation by streptokinase (SK), we expressed the cDNAs corresponding to each SK domain individually (α, β and γ), and also their two-domain combinations viz. αβ and βγ in E. coli. After purification, α and β, showed activator activities of around 0.4 and 0.05 %, respectively as compared to that of native SK, measured in the presence of human plasmin, but the bi-domain constructs αβ and βγ showed much higher co-factor activities (3.5 % and 0.7 % of native SK, respectively). Resonant Mirror-based binding studies showed that the single-domain constructs had significantly lower affinities for 'partner' HPG whereas the affinities of the two-domain constructs were remarkably native-like with regards to both binary-mode as well as ternary-mode ('substrate') binding with HPG, suggesting that the vast difference in co-factor activity between the two-and three-domain structures did not arise merely from affinity differences between activator species and HPG.
Remarkably, when the co-factor activities of the various constructs were measured with µPG, the nearly 50-fold difference in the co-factor activity between the two-and three-domain SK constructs observed with full-length HPG as substrate was found to be dramatically attenuated, with all three types of constructs now exhibiting a low activity of around 1-2 % compared to that of SK.HPN and HPG. Thus, the docking of substrate through the catalytic domain at the active site of SK-plasmin(ogen) is capable of engendering, at best, only a minimal level of co-factor activity in SK.HPN and further gain in this basal co-factor activity to native-like levels is brought about via kringle domains likely at a post substrate-docking step. Therefore, apart from conferring additional substrate affinity through kringle-mediated interactions, reported earlier (Dhar et al., 2002; J. Biol. Chem. 277: 13257), selective interactions between all three domains of SK and the kringle domains of substrate vastly accelerate the plasminogen activation reaction to the levels associated with the native activator species.

INTRODUCTION
Streptokinase (SK) 1 is a widely-used bacterial thrombolytic protein that is secreted by several species of β-hemolytic streptococci (1,2). It consists of a single polypeptide chain of 414 residues and is organized into three structurally similar, independently folding domains (termed α, β and γ in the order from N-to C-terminus of the polypeptide), that are separated by coiled coils and small flexible regions at the two ends (3)(4)(5). Like several other well-known thrombolytic proteins, such as urokinase (UK) and tissue-plasminogen activator, SK exerts its effects through the conversion of human plasminogen (HPG) to its proteolytically active form, plasmin (HPN). Thus, during the treatment of various circulatory disorders e.g. myocardial infarction, deep vein thrombosis, pulmonary embolism etc, HPN generated by the activation of HPG helps restore blood flow to the afflicted part by proteolytic dissolution of the fibrin in the pathological clot. In contrast to tissueplasminogen activator and UK, which are intrinsically HPG-specific proteases and thus 'directly' act on HPG, SK is an enzymatically inert protein (6). However, it recruits circulating HPG to generate HPG activating potential; SK first combines with 'partner' HPG in an equimolar manner to form a tight, enzymatically active complex, the so-called SK.HPG 'virgin' activator complex (7,8) which rapidly converts into a SK.HPN complex. The mature SK.HPN activator complex then catalytically transforms 'substrate' molecules of HPG to HPN (1).
whereby SK manages to switch the substrate specificity of the active site of HPN (5), indications of which had also been gleaned earlier through solution studies (14,15). These studies indicate that SK acts as a protein co-factor of the plasmin active site by forming a 'three-sided valley' around the active center by virtue of its tri-domain structure. Thus, it has been postulated that the docking of the catalytic sub-unit of HPG onto the activator complex by protein-protein interactions facilitates cleavage of the scissile peptide bond in the macromolecular substrate by an otherwise non-specific active site (5). However, the crystal structure does not provide any indications whether the kringle domains of the substrate play any role in this process.
In contrast to the three-domain structure of SK, staphylokinase (SAK), another functionally similar, 'indirect' HPG activator protein, but one that has a single domain only, also operates by docking substrate HPG into the active center of its binary complex with plasmin(ogen) (16). Although SAK displays little sequence similarity with SK, its single domain bears a remarkable conformational similarity with one of the three SK domains in particular (the α domain), and all the three domains of SK in general, which, in turn, are structurally homologous to each other (17). However, the SAK-HPN activator complex is known to possess a significantly lower catalytic efficiency than SK-HPN (18,19). Apart from the single-domain prototypic structure of SAK, another mechanistically similar plasminogen activator from Streptococcus uberis (abbreviated SUPA) has also been isolated, which possesses a two-domain structural motif (20,21). Thus, a compelling need exists to glean insights regarding the similarities and dissimilarities in the 'design principles' between the three-domain SK, on the one hand, and the single-and two-domain structures of SAK and SUPA on the other, particularly in terms of the structure-function co-relations that underlie the inter-domain co-operativity between the individual SK domains.
Taking a minimalist approach, a pertinent question that can be posed in the above context is whether any of the isolated domains of SK also possess, like the single domain of SAK, the ability to bind with HPG in both substrate and partner modes and, if the answer is affirmative, whether this binding is functionally translated into a capability, even if highly compromised compared to native SK, to switch the non-specific substrate preference of partner plasmin to that of a HPG activator enzyme. A recent study (22) has attempted to address this issue, particularly in terms of the functional properties of single SK domains, and elegantly demonstrated the presence of detectable, albeit very low levels of HPG activator activity in two of the three SK domains.
The results of this study also stimulate curiosity about the possiblity of a hierarchical generation of catalytic activity in SK (i.e. the possible existence of a progressive increase in co-factor activity from the single-to two-domain, and finally, to the three-domain native motif), and the need to gain meaningful insights into the design principle/s selected by evolution for plasminogen activators of bacterial origin. As a first step in this direction, in the present study, therefore, we have expressed each of the three domains of SK, as well as their two-domain combinations (viz. αβ and βγ) in a heterologous expression system, and compared their catalytic and binding properties both against full-length macromolecular substrate, HPG as well as its kringle-less derivative, micro-HPG (µPG). Thus, with this approach, we have attempted to explore the interdomain co-operation prevalent in SK, as well as between the SK.HPG activator complex and substrate HPG. The results, presented below, reveal that the kringle domains of substrate help not only in substrate docking, as proposed recently (23,24), but also in substrate turn-over, even though they are distinctly apart from the immediate microenvironment of the 'target' scissile peptide bond --the specific region of the substrate polypeptide that is selectively cleaved during plasminogen activation.

EXPERIMENTAL PROCEDURES
Materials -HPG was either purchased from Roche Molecular Biochemicals (Mannheim, Germany) or purified from human plasma by affinity chromatography (25). The cloning vector pBluescript II KS -, thermostable DNA polymerase (pfu) and E. coli strain XL-Blue were procured from Stratagene Inc. (La Jolla, CA). All other enzymes used for genetic manipulation were obtained from New England Biolabs (Beverly, MA). Oligonucleotide primers were either synthesized in-house on an Applied Biosystems DNA synthesizer model 492 or customsynthesized by Ransom Hill Biosciences Inc. (Ramona, CA). DEAE-Sepharose (Fast-flow) and Chelating-Sepharose were procured from Pharmacia Biotech. Ltd., Uppsala, Sweden, and Phenylagarose for hydrophobic interaction chromatography (HIC) was purchased from Affinity Chromatography Ltd., Isle of Man, U. K. The T7 RNA polymerase promoter-based expression vector, pET-23d and Bug buster ® (a commercial reagent for rapid bacterial cell lysis) were procured from Novagen Inc. (Madison, WI). All other chemicals used were of the best commercial grade available.
Expression and purification of SK from E. coli -The plasmid construct for the intracellular expression of SK in E. coli (pET-23d-SK) under the control of the T7 phage RNA polymerase promoter has been described (26). Briefly, the purification involved lyzing the cells by sonication followed by ammonium sulfate precipitation (60 % saturation) and two chromatographic steps viz. HIC on phenyl-agarose and anion exchange chromatography. This yielded homogeneous SK with an overall yield of 55-60 % with a specific activity of 1.1 x 10 5 I.U./mg protein (27).

Cloning of single and bi-domains of SK -
The exact boundaries of individual domains were selected based on limited proteolysis data (4), which also cross-correlated well with the limits defined by the crystal structure of SK (5). The sequences 1-143 and 143-293 were chosen to represent the α domain and the β domain, respectively (4). The γ domain (SK293-414) was cloned from the region coding for residues 300-387, since the C-terminal end of the γ domain (residues 388-414) contains residues that are not critical for the functioning of full-length SK and forms a disorganized and flexible segment that undergoes rapid degradation in the presence of HPN (4,5,28,29).
The construction of the truncated derivatives of SK was carried out by PCR amplification of the desired sequence, using SK as template, and specific upstream and downstream primers which also contained specific RE sites to allow for facile, directional Purification of β domain of SK -The β domain of SK (residues 143-293) was expressed in E. coli as a soluble protein at a level approximately 30 % of the total soluble protein fraction, and was purified as described earlier (23).
Purification of the two-domain constructs αβ (SK 1-293) and βγ (SK 143-414) -The covalently contiguous two-domain construct, αβ, was expressed as a soluble protein in E. coli BL-21 cells, whereas the βγ protein was expressed predominantly as IBs. The purification involved lyzing the cells by sonication followed by ammonium sulfate precipitation (60 % saturation) and two tandem chromatographic steps, namely hydrophobic ion chromatography on phenyl-agarose and ion exchange chromatography on DEAE-Sepharose (Fast-flow) essentially as described for SK earlier in the Experimental Procedures. The eluted fractions contained protein that was more than 95 % pure as analyzed by SDS-PAGE.
Measurements were carried out from 190-250 nm in a 0.1cm path length cuvette, and the appropriate buffer baseline was subtracted from the protein spectra. The final spectrum analyzed was an average of 10 scans. Secondary structure analysis to compute the content of alpha helix, beta sheet, random coil and other secondary structure/s were carried out using the algorithm described by Yang et al., 1986 (31).
Preparation of microplasminogen (µPG) -Microplasminogen, the catalytic domain of plasminogen (residues Lys 539 -Asn 790 ), devoid of all the kringles was prepared by cleavage of HPG by HPN under alkaline conditions (0.1N glycine/NaOH buffer, pH 10.5, 10:1 ratio of HPG and HPN) at 30° C. Microplasminogen was purified from the reaction mixture by passing through a Lys-sepharose column (Amersham Biosciences), followed by a Soyabean-trypsin-inhibitor (STI)-Sepharose CL-4B column to absorb HPN and µPN, as reported (23,32). The flow-through was then subjected to molecular sieve chromatography, after concentration by ultrafiltration, on a column (16 x 60 cm) of Superdex-75 TM (Amersham). The purity of µPG formed was analyzed by SDS-PAGE, which showed a single band moving at the position expected from its molecular size (32), and the identity of the protein was confirmed by N-terminal protein sequencing. Activation with urokinase, which is known to be an efficient activator of µPG despite the absence of kringle domains in the latter, was used to establish that the rate of activation of this preparation, when used as substrate, was comparable to that obtained when using SK.HPN as the activator species.
Assay for detection of extremely low HPG activation capability -Catalytic amounts of domains were added to the assay buffer (100 mM Hepes, pH 7.5) containing 2 µM of HPG and chromogenic substrate (0.5 mM) and the change in absorbance was monitored spectrophotometrically at 405 nm as a function of time at 22°C (Pathway I). In order to detect intrinsically low HPG activation capability in single-domain/bi-domain construct/s that may otherwise not be capable of activating the partner HPG unless provided with a preformed active site (i.e. only Pathway II capability), equimolar complexes of HPN and the SK derivative were premixed on ice for 1 min, and catalytic amounts were withdrawn and added to 2 µM HPG and chromogenic peptide substrate in the assay buffer containing 100 mM Hepes, pH 7.5, in case of single-domain proteins and 50 mM Tris-Cl, pH 7.5, in case of the two-domain proteins. The reactions were recorded spectrophotometrically at 405 nm and the progress curves obtained from corresponding control reactions (containing HPG and same amounts of HPN, but no domain) were subtracted from the test reactions. The resultant curves so obtained were used to determine the rate of HPG activation, which was then expressed as % activity, relative to SK. The initial phase (0-8 min) was generally used for this calculation in both cases, to overcome any loss of activity at later stages due to possible proteolytic stability.
In order to establish whether the isolated domains had genuine activator activities over and above that associated with the background cells, E. coli BL21 cultures transformed with the pET-23d vector but not containing any domain-encoding insert, after induction with IPTG, were harvested and lyzed similarly as ones prepared for purifying the individual domain/s. These cell lysates were subjected to exactly the same treatments as the ones prepared from cultures harboring plasmids for α, β or γ, including immobilized metal affinity chromatography (IMAC) and desalting steps. The fractions from 'control' columns (on which the lysates containing just the expression plasmids without any SK domain insert were processed) identical in volume to the corresponding eluted protein fractions from the 'test' columns (from which the domains were being purified) were collected in parallel and these were then used for assessing HPG activator activity. The 'control' fractions, which served to measure the background E. coli cell-associated HPG activator activity consistently showed a very low level of activity (< 0.005 %) but one that was significantly lower than the activity observed with any of the truncated domain constructs.
Determination of kinetic constants for HPG activator activity -Varying amounts of HPG were added to the assay cuvette containing fixed amounts of single-domain or bi-domain constructs and chromogenic substrate (0.5 mM) and the change in absorbance was monitored at 405 nm as a function of time at 22°C. Also, the kinetics of HPG activation by the activator complexes were measured by transferring suitable aliquots of preformed complexes (single-domain or bi-domain constructs, and HPN) to the assay cuvette containing different concentrations of substrate HPG.
To compute the k cat , the number of functional HPN active sites was determined by titration with an active-site acylating reagent, NPGB (7,33,34).

Kinetic Analysis of Protein-Protein Interactions using Resonant Mirror Technology:
Binary interaction analysis-Association and dissociation between HPG and the nSK/SK domains viz. bi-domains (αβ and βγ) and single-domains (α and β) referred to hereafter as binary interaction, were followed in real time by resonant mirror-based detection using IAsys Plus TM system (Cambridge, UK) (35,36). In these experiments, streptavidin was captured on biotin cuvettes. This was followed by the attachment of biotinylated HPG to the streptavidin captured on the cuvette. Nonspecifically bound HPG was then removed by repeated washing with phosphate-buffered saline followed by three washes with 10 mM HCl. The net response choosen for the immobilized biotinylated HPG onto the cuvette was 700-800 arc seconds in all the experiments. Experiments were performed at 25 0 C in 50 mM phosphate buffered saline, pH 7.4 containing 0.1% Tween 20, 250 mM NaCl, and 50 µM NPGB (binding buffer). The latter was included in order to prevent plasmin-mediated proteolysis.
After equilibrating the cuvette with binding buffer, varying concentrations of either nSK/SK-domains were added, and each binding response was monitored during the "association" phase. Subsequently, the cuvette was washed with the binding buffer, and the "dissociation" phase was recorded (37). Following each cycle of analysis, the cuvette was regenerated by washing with 10 mM HCl, and base line was re-established with the binding buffer. In parallel, in the control cell in the dual channel cuvette, immobilized streptavidin alone was taken as a negative control for the binding studies.
The data were analyzed after subtraction of the corresponding non-specific refractive index component(s), and the kinetic constants were calculated from the sensorgrams by nonlinear fittings of association and dissociation curves using the software FASTfit TM , supplied by the manufacturer. The dissociation rate constant (k d ) was calculated from the average of four dissociation curves obtained at saturating concentration of ligate. The equilibrium dissociation constant (K D ) was than calculated from the extent of association of monophasic curve. The k a was calculated from the equation k d /K D. Values of K D obtained using this relationship were in good approximation to those obtained by k d /k a obtained from the linear fit of k on versus ligate concentration (38).
Ternary Interaction Analyses-Resonant mirror technology-based biosensor was also used to measure the rate and equilibrium dissociation constants describing interactions between soluble ligates (HPG and µPG) and nSK/SK-domains complexed with immobilized HPG. In binary interaction studies, it was evident that when soluble nSK/SK-domain was added to the immobilized HPG, a rapid and avid nSK/SK-domains.HPG (except single domains, α and β) binary complex formation occurs. The dissociation of these binary complexes is very slow due to the high stability of SK.HPG complex (23). After allowing the complex to dissociate maximally (~20 min), the dissociation baseline becomes stable, which remains unaffected even after washing with 2.5mM EACA. In case of single domains (α and β), weak binary complexation with immobilized HPG was observed, and it was seen that the dissociation rates in binding buffer itself were very high; thus, stable baselines as observed in case of bi-domains and nSK were not observed.

Isolation and characterization of isolated single and bi-domains of SK:
In order to measure the HPG activator activities in isolated domains and bi-domains of SK of S. equisimilis, each of the three isolated domains (α: SK1-143; β: SK143-293; γ: SK300-387) and bi-domain constructs (αβ and βγ) were constructed by PCR amplification of cDNAs encoding for respective domains followed by cloning in pET-23d, a T7 phage RNA polymerase promoter-based vector (26). All of these constructs were then expressed in E. coli BL21 (DE3) cells. Good expression levels were observed in case of all of these constructs, and purification strategies for each of these were then standardized (see Experimental Procedures) to obtain preparations with purity levels well over 95 % as judged by SDS-PAGE (Fig. 1A). The secondary structure content of the isolated single-and double-domain derivatives was determined using far-UV CD measurements (Fig. 1B.). The secondary structural features of the SK derivatives indicated that these were folded, as these were found to be largely in concordance with the reported data in literature, either based on spectroscopic analysis (4,28), or that expected from direct X-ray structural data on SK domains (5,39,40).
Substrate HPG activation by SK can proceed along two Pathways (41) -Pathway I, wherein, after a high affinity complexation occurs between SK and partner HPG molecule, a 'virgin' site is formed in the zymogen in absence of any proteolytic cleavage (7,8). However, this complex rapidly converts to SK.HPN, the activator species that catalytically activates molecules of substrate HPG, after the partner HPG has been proteolytically converted to plasmin. In Pathway II, SK can directly complex with active HPN and then can activate substrate HPG.
Mutants of SK that are impaired in Pathway I capability can sometimes activate substrate HPG through pathway II if 'supplied' with a preformed, partner active site in the form of HPN (22).
Thus, zymogen activation of partner HPG by SK under Pathway I, and the conversion of macromolecular substrate specificifity of the HPN active site upon the latter's complexation with SK, constitute two distinct (although partially overlapping) biochemical properties of HPG activation with this activator protein.
We first examined the functional ability of purified individual (single) domains to activate substrate HPG using either excess HPG (Pathway I) or with excess substrate HPG after pre-mixing with equimolar amounts of HPN (Pathway II). The γ-domain did not show any detectable activator/co-factor activity under either of these assay conditions, but in case of α and β domains, apparent activator activities that could be detected using HPG alone were not more than 0.002 % as compared to the corresponding activity with SK, taken as 100 %. These results are very similar to those reported by Loy et al (2001). However, the barely perceptible activator activity associated with α and β were found to be strongly HPN dependent, since addition of small amounts of HPN in the assay mixtures resulted in sharp enhancement of HPG activation rates (data not shown); indeed when activator activity of α domain against substrate HPG was measured after premixing with HPN, it was found to increase dramatically by several orders of magnitude compared to its activity measured directly with HPG. The α domain now exhibited an activity approximately 0.5 % that of native SK, while the β domain also showed enhanced activity, of around 0.05 % compared to SK (see Table I, which shows the steady state kinetic data for HPG activation). However, it is clear that despite the availability of preformed HPN as partner, the co-factor activities of both the single domains remained, at best, a small fraction of that observed with native nSK.
The presence of relatively low co-factor activities in the isolated domains prompted us to examine if two-domain combinations of SK acquired a greater capability for catalysis as compared to the single-domain constructs. The latter (αβ and βγ) were therefore tested for their ability to activate substrate HPG in a HPN-independent or -dependent manner, as before. It was observed that, like the case of single-domains, αβ and βγ had very low activities in the absence of HPN (< 0.05%), as compared to that of nSK, but in presence of HPN, significantly higher catalytic ability as compared to that seen with individual domains could be observed; αβ and βγ had around 3.5 % and 0.7 % co-factor activities in presence of HPN, respectively, as compared to that of nSK measured under similar conditions (see also Fig. 2 for comparative HPG activation progress curves of single-and two-domain constructs). The steady-state kinetic constants obtained with HPG as substrate for two-domain constructs are also shown in Table I along with those of single domains, all activities being measured after making 1:1 complexes with HPN prior to the assay in presence of excess substrate HPG. It may also be mentioned that equimolar mixtures of either single domains, or that of mixtures of two, or three domains together did not result in any detectable enhancements in activities over those additively associated with the individual domains, indicating that the covalent contiguity between the individual domains in αβ and βγ contributed synergistically towards the acquisition of higher activities by the two-domain construct/s. An analysis of the steady-state kinetic parameters shown in Table I indicates that although the two-domain molecules were minimally altered in K m values for HPG (around 0.5 µM in case of nSK), they still exhibited a significantly reduced k cat for HPG activation when compared to the native protein. In contrast, we observed a nearly 3-4 fold increase in K m (~ 2 µM) for substrate HPG binding to activator complex in case of the single domain, α, thereby indicating that the isolated domain had probably lesser affinity for substrate HPG in comparison to both full-length native nSK and the bi-domain derivatives. Although these data, together, clearly establish that there is a dramatic increase in the co-factor capability as the single domain/s of SK acquire a two-domain character, the latter are still significantly compromised in terms of their catalytic power when compared to the three-domain, native format.

Interaction of single and bi-domain constructs of SK with HPG:
The inability of single-and bidomain derivatives of SK to activate substrate HPG, even with HPN, beyond a certain low level prompted us to exanmine their capability (or lack thereof) to bind HPG. For this purpose, various real-time binding parameters such as association and dissociation rate constants and equilibrium dissociation constants for binary interactions between immobilized biotinylated-HPG and nSK/SK-domains were first determined by the Resonant Mirror approach using a real-time molecular interaction system (IAsys Ltd., U.K.). The kinetic parameters so obtained for the binary interaction between immobilized HPG and nSK, or those of single-(α and β) and bidomain constructs (αβ and βγ) are shown in Table II. It is clear from these results that the equilibrium dissociation constant (K D ) of αβ (~2 nM) and βγ (~7.5 nM) are not grossly different from that of nSK (~1.0 nM), while the K D values of single domains (~100 nM) are nearly two orders of magnitude greater than that for nSK. These observations indicate that partner HPG binds to the bi-domains with nearly the same affinity as it binds to nSK, whereas single domains bind to partner HPG with considerably lesser affinity as compared to native nSK. It is thus possible that the lowered affinity between the single domain/s and HPG may be a major contributive cause to their overall lowered co-factor activity. However, the native-like affinity observed for HPG in case of the two-domain constructs, particularly αβ, suggests that the underlying cause of their lowered co-factor activities does not likely reside at this level of intermolecular interaction.
After establishing the stability of binary complexes of the different constructs with immobilized HPG, ternary interactions between 'substrate' HPG and different domains of nSK (single-and bi-domains) pre-complexed with immobilized HPG were explored using, again, an approach based on the Resonant Mirror technique. The application of the Resonant Mirror technique to detecting and studying ternary interactions in the SK.HPG system has been reported recently (23). In this 'direct' approach for analyzing the protein-protein interactions operative between macromolecular substrate and the SK.HPG activator complex, mildly biotinylated HPG is first immobilized onto the surface of a Resonant Mirror biotin-cuvette onto which a thin film of streptavidin has earlier been layered, followed by washing with buffer to remove excess unbound HPG. The immobilized HPG is then 'challenged' with SK to form equimolar SK.HPG complex.
The binary complex so formed is washed extensively with buffer, a procedure against which the high-affinity SK.HPG binary complex is fairly stable (23). To the binary complex is then added excess 'substrate' HPG; the formation of a ternary HPG.SK.HPG complex can then be observed to give additional 'ternary signal' over and above that obtained earlier at the binary complexation stage, as shown in Fig. 3 (see also ref. 23). Thus, this assay, which is capable of monitoring in real-time, the 'docking' of substrate HPG onto the pre-formed SK.HPG binary complex can be utilized to explore altered substrate-interacting capability, if present, in a SK derivative/mutant after prior complexation with partner HPG. In the past, we have successfully employed this approach to demonstrate affinity changes between the SK-plasmin(ogen) activator complex and its altered form for the macromolecular substrate, HPG (23). When this assay was carried out in the present study (see "Experimental Procedures" for details), it was observed that equilibrium dissociation constant (K D ) of the ternary interaction of substrate HPG with bi-domains of nSK (viz. αβ and βγ) was similar to that of native SK for substrate HPG (~0.12 µM). The association rate constant (k a ) and dissociation rate constant (k d ) of ternary interactions between both the bidomain of nSK and substrate HPG were found to be nearly similar to that observed in case of ternary interaction of substrate HPG with native full length nSK (Tables I and II). However, in case of the single domains (α and β), real-time kinetic constants for ternary complexation could not be determined at all due essentially to the weak nature of binary complex formation, in the first place, between the isolated domain/s and immobilized HPG (Fig. 3, panel C). This was so because, in this case, although the binary complex/es could be observed when much higher HPG concentrations than those required for nSK were employed, these were seen to dissociate rapidly even by moderate buffer washing -conditions under which the native SK.HPG was stable. In contrast to single domains, however, the results of Resonant Mirror studies on αβ and βγ clearly demonstrate that these seem to be as capable of docking macromolecular substrate as the native tri-domain nSK molecule.
Taken together, the steady-state kinetics and real-time physico-chemical studies of substrate HPG interaction with nSK/SK-domains clearly establish that although substrate HPG docks with both the bi-domain derivatives of SK with nearly the same affinity as that with fulllength SK, yet this near-native docking of substrate HPG is not proportionately translated into a near-native co-factor activity in either of the two bi-domain derivatives, and the native threedomain structure is catalytically the most effective in terms of catalytic turnover. The lack of any significant quantitative co-relation in a given derivative between the docking ability of substrate, on the one hand, and its subsequent catalytic turnover, on the other, strongly indicated that these two attributes likely represent two independent phenomena. Therefore, it appeared probable that there exist specific interactions between the macromolecular substrate and the SK.HPN activator complex at the post-docking stage that play important role/s in amplifying the relative low catalytic power associated with the two-domain constructs to that associated with the native tridomain nSK --ones that are not being effectively utilized by the bi-domains despite a near-native ability to dock the macromolecular substrate. If this were indeed true, one can further extend this model to assume that these interactions may either be short-range i.e. localized in and around the scissile peptide bond region in the catalytic domain of substrate HPG, or these could be relatively long-range, such as those involving the kringle domains of substrate. Earlier studies on SK have provided strong indication that kringle domains of substrate HPG do play an important role in HPG activation by nSK, since µPG, which is devoid of all the five kringle domains of HPG has both significantly lowered affinity for the preformed SK.HPN activator complex (23) and is a poor substrate for activation (32). Thus, it is conceivable that both catalytic domain-centred as well as kringle domain-related interactions are operative together during substrate recognition and turn-over by the SK.HPN activator complex. Hence, with a view to obtain insights into the relative contributions of different domains of substrate HPG (i.e. catalytic versus kringle domains) we compared the rates for activation of HPG by single-, bi-and tri-domain structures using as substrate, with those for the isolated catalytic domain, µPG (which also, like native fulllength HPG, contains the scissile peptide bond but is devoid of all the kringle domains).
Interaction of µPG with nSK/SK-domains: Microplasminogen was generated from HPG by enzymatic cleavage with plasmin (see "Experimental Procedures" for details). Kinetic experiments were then performed using µPG as substrate and the different HPN-SK domain/s as the activator species. The results demonstrated that in contrast to the activator activity of native SK.HPN with full length HPG as substrate (taken as 100 %), all the three types of SK structures (single-, bi-and tri-domain) showed significantly low specific activity, ranging roughly over 1.0-2 % when µPG was used as the substrate (see Table III for steady-state kinetic data).
Interestingly, in general the K m of nSK/SK-domains for µPG was seen to be 4-5 fold lesser than that for full-length HPG, which, at least partially accounts for the generally decreased specific activities observed when µPG was the substrate. However, these low activities against µPG could not be explained merely on the basis of their lowered affinity for the substrate since even at substrate saturation the activities of their HPN complexes remained essentially 40-50 fold lower as compared to that of SK.HPN against full-length HPG (Fig. 4). These results clearly establish that the co-factor activity of the various constructs against the kringle-less substrate could not be 'compensated' by saturating the assays with substrate. Thus, the single-and two-domain constructs were likely truly compromised not just in terms of the substrate affinity but in terms of their catalytic power as well. It is therefore apparent that when kringles are 'lost' from the macromolecular substrate, the catalytic advantage of even the native, three-domain SK becomes compromised, showing that the catalytic domain of macromolecular substrate is intrinsically a 'poor' substrate for the SK.HPN activator complex. Another noticeable feature of the results is that in case of both the bi-domain derivatives of SK, the K m for µPG was found to be nearly same as that of SK (~2.0 µM), while in case of the α domain there was an approximately 2-fold difference in K m for µPG as compared to that of native SK (Tables II and III). This observation indicates the presence of additional, catalytic domain-specific substrate affinity in the two-domain constructs as compared to the single-domain derivative, α. The results also reveal that with µPG as substrate, the catalytic activity of the bi-domain construct αβ (although exhibiting the generally lowered activity of all the constructs, including native SK, against the kringle-less substrate) was nearly two-fold lower as compared to its activity against HPG (see Tables I and   III). This suggested that unlike the single domains, the two-domain constructs possessed an improved capability to interact with the kringles of substrate, resulting in a certain degree (although modest compared to the three-domain design) of catalytic enhancement, even though other indicators of enzyme-substrate interactions, such as K m for substrate and direct, real-time affinity measurements using Resonant Mirror analyses did not yield any definitive evidence providing a physico-chemical basis for this functional effect. However, it is clear from the steadystate kinetic studies that the binary complexes of nSK/SK-domains (i.e. the tri-and bi-domains) interact with substrate µPG with essentially similar K m , although in case of the α domain, the values are somewhat higher, indicating a relatively lowered substrate affinity. Remarkably, in all cases, even at saturating µPG concentrations, only poor co-factor activities compared to that of SK.HPN against full-length HPG were engendered. In order to directly validate these conclusions, real-time binding parameters of ternary interactions between µPG and nSK/SKdomains complexed to immobilized biotinylated-HPG were determined using the Resonant Mirror technique. The data (Table III), which are broadly consistent with the kinetics' results, show that equilibrium dissociation constants (K D ) of ternary interaction between µPG and αβ (~4.0 µM) and βγ (~2.0 µM) were, at most, minimally different compared to that for nSK (~1.0 µM). Thus, these analyses suggest that the binary complexes of the bi-domains and the native, tridomain nSK with HPG dock µPG with affinities that are not grossly different from each other.
However, it is worth noting that as compared to the docking of substrate HPG onto the binary complexes of nSK/SK-domains, µPG docks onto the latter with an affinity that is approximately an order of magnitude lower.
One of the main conclusions emerging from the present study is that kringle domains of HPG play an important role in docking of the macromolecular substrate onto the native activator complex. However, although the macromolecular docking phenomenon is crucially important in generating nascent co-factor activity in the HPN active site, it is probably not the sole determinant of the 'full-blown' co-factor activity associated with the native system viz.
SK.HPN versus HPG as the substrate. This conclusion is supported by the observation that both the two-domain derivatives, αβ and βγ can dock HPG nearly as efficiently as the native activator complex (as indicated by their respective K m and K D values), but fail to generate native-like cofactor activity. In other words, the docking of substrate by the two bi-domain structures is nativelike, yet their catalytic rates are a mere fraction of the rates seen with native SK and full-length HPG. Thus, it seems reasonable to conclude that it is only when all three domains of SK simultaneously present that the ability to exploit the kringle domains for improved catalytic rates (as opposed to 'mere' improvement in substrate affinity) effectively gained. Thus, kringles help not only in docking substrate at the active site, but also dramatically accelerate the rates of catalysis by the activator complex. The present studies strongly suggest that events subsequent to substrate docking, which require all three domains of native SK together with the kringle domain(s) of the substrate, are critical in the development of the full-blown catalytic activity of the SK.HPN activator enzyme, and although stereo-chemical positioning of the scissile peptide bond is necessary in conferring a native-like macromolecular specificity, it can generate limited catalytic turn-over on its own, with the major jump in catalytic power being achieved through long-range protein-protein interactions with substrate kringles. This mechanism of substrateassisted proteolysis encountered in SK is clearly distinct from that of other 'direct' HPG activators like TPA and UK, which display catalysis with nearly the same rates irrespective of the presence or absence of kringles in the substrate 2 . However, of all these activators (including SAK), the catalytic power of the SK.HPN complex is recognized to be the highest (18,34,42), probably due to the recruitment of long-range interactions between enzyme and its protein substrate in its catalytic mechanism of action. This may indeed be the underlying structural reason for the evolutionary selection of a three-domain activator design, particularly one in which specific interactions between the activator species and macromolecular substrate are centered not only on the 'target' of proteolysis, the catalytic domain (which, by itself, is not surprising) but also with regions located far away from the site of enzymatic action by the active site per se.
Although speculative at present, it seems plausible that the tri-domain molecule is able to optimally 'sense' a signal during catalysis, probably conformational changes in the catalytic domain as the scissile peptide bond is cleaved, as well as the relatively large structural changes that are known to occur in the orientation of the kringles following the conversion of substrate plasminogen to plasmin (43,44). This process, by facilitating the release of the product of the catalytic cycle from the active site, would help attain the native-like protein co-factor activities seen in the SK system. Further detailed studies are undoubtedly required now to elucidate the exact mechanism, for example those exploiting site-specific spectroscopic monitoring for FRET analysis to help probe the molecular motions during substrate capture, catalytic transformation and, finally, product release. However, the results presented in this study, by illuminating the initial steps in an otherwise extremely complex biochemical reaction, unambiguously indicate that the substrate-specific proteolytic capability generated in the immediate vicinity of the HPN active site upon complexation with SK is amplified in a dramatic manner through long-range, 'supra-catalytic center based' protein-protein interactions between the co-factor and macromolecular substrate. The elucidation of the exact molecular events, and epitopes, involved in this process would greatly help in the redesign of existing proteases into efficient substrate specific HPG activator enzymes, and also in the future de novo design of novel target-specific proteolytic functionalities with useful applicabilities.  were added onto the binary complexes. However, for clarity in the figure, only a single (saturating) concentration of HPG (1 µM) is depicted (shown in panels A and B by arrows, marked 3). The association phases were monitored for ~5 min and subsequently the cuvettes were washed with the binding buffer (point of addition depicted in panels A and B by vertical arrows marked 4). After each cycle of analysis in A and B, the undissociated substrate HPG was stripped off with 2.5 mM EACA, followed by re-equilibrating the cuvette with binding buffer, which reestablished the original baseline (data not shown). The non-specific binding interaction between added substrate HPG with the HPG immobilized onto the cuvette in the absence of any SK/SKderivatives is depicted in panel D (arrow 1 depicts the addition of binding buffer, arrow 2 depicts the addition of HPG and arrow 3 depicts the dissociation of non-specific HPG with the binding buffer). Note that the non-specific signal was found to be less then 10% of the ternary signal in case of SK/SK-derivatives. In a separate experiment the immobilized streptavidin alone was also taken as a negative control, and it was subjected to the same kind of treatments as given to the test cell containing immobilized HPG. Under these conditions no significant non-specific binding was observed. Detailed protocols and rationale for these experiments are provided under "Experimental Procedures" and in ref. 23.

FIG. 4. Comparison of substrate HPG and µPG activation capability of nSK/SK-domains.
The histogram depicts the comparison of cofactor activity of SK, bi-domain constructs of SK (αβ and βγ), and single domain constructs (α and β); in all cases, activity was measured alongwith equimolar HPN, for the activation of either of two macromolecular substrates i.e. HPG, and µPG.
The saturating concentration of HPG used in case of tri-(nSK) and bi-domains (αβ and βγ) is 2 µM while in case of single-domains (α and β) the co-factor activity is measured at 8 µM HPG.
The cofactor activity of tri-(nSK) and bi-domains (αβ and βγ) was measured at saturating concentration of µPG (4 µM) while cofactor activity of single-domains (α and β) was measured at 10 µM µPG. The percentage activity of different activator complexes was then calculated as a percentage value in comparison to the co-factor activity of SK.HPN when HPG was the substrate taken to be 100%. The histogram shows maximal percentage of co-factor activity of different domains of SK against HPG (black) and µPG (light grey) at their respective saturating substrate concentrations. The data highlights the fact that of all the various constructs and substrate types, only the tri-domain motif is able to generate the full (100 %) activity associated with native SK and full-length HPG, while all the other partial-length constructs with either substrate, or, even The kinetic parameters for co-factor activity against HPG as substrate were determined with the respective activator complexes at 22 0 C as described under "Experimental procedures". The data represent the mean of three independent determinations.

ND:
Could not be accurately determined due to the intrinsically low level of co-factor activity.

Table II
Association and dissociation rate constants and apparent equilibrium dissociation constants of 'partner' and 'substrate' HPG with SK/SK-domains.