Biochimica et Biophysica Acta 1804 (2010) 1730–1737

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r.com/ locate /bbapap

Probing the primary structural determinants of streptokinase inter-domain linkersby site-specific substitution and deletion mutagenesis

Suman Yadav, Girish Sahni ⁎Institute of Microbial Technology (C.S.I.R), Sector 39-A, Chandigarh-160036, India

Abbreviations: EDC, N′-ethylcarbodiimide; HPG, humplasmin; IBs, inclusion bodies; IPTG, isopropyl-1-thequilibrium rate constant; koff, rate of dissociation;microplasminogen; NHS, N-hydroxysuccinimide; NPGBbenzoate; nSK, SK derived from S. equisimilis strain H46splicing overlap extension-polymerase chain reaction; SPSTI, soybean trypsin inhibitor; WT-SK/rSK, SK of S. equi⁎ Corresponding author. Division of Protein Science

Microbial Technology, Sector 39-A, Chandigarh-160036,fax: +91 172 2690585.

E-mail address: sahni@imtech.res.in (G. Sahni).

1570-9639/$ – see front matter © 2010 Elsevier B.V. Adoi:10.1016/j.bbapap.2010.04.003

a b s t r a c t

a r t i c l e i n f o

Article history:Received 9 October 2009Received in revised form 10 February 2010Accepted 4 April 2010Available online 24 April 2010

Keywords:Enzyme–substrate interactionLinkerInter-domain interactionPlasminogen activatorStreptokinase

The bacterial protein streptokinase (SK) contains three independently folded domains (α, β and γ),interconnected by two flexible linkers with noticeable sequence homology. To investigate their primarystructure requirements, the linkers were swapped amongst themselves i.e. linker 1 (between α and βdomains) was swapped with linker 2 (between β and γ domains) and vice versa. The resultant constructexhibited very low activity essentially due to an enhanced proteolytic susceptibility. However, a SK mutantwith two linker 1 sequences, which was proteolytically as stable as WT-rSK retained about 10% of theplasminogen activator activity of rSK When the native sequence of each linker was substituted with 9consecutive glycine sequences, in case of the linker 1 substitution mutant substantial activity was seen tosurvive, whereas the linker 2 mutant lost nearly all its activity. The optimal length of linkers was then studiedthrough deletion mutagenesis experiments, which showed that deletion beyond three residues in either ofthe linkers resulted in virtually complete loss of activator activity. The effect of length of the linkers was thenalso examined by insertion of extraneous pentapeptide sequences having a propensity for adopting either anextended conformation or a relatively rigid conformation. The insertion of poly-Pro sequences into nativelinker 2 sequence caused up to 10-fold reduction in activity, whereas its effect in linker 1 was relativelyminor. Interestingly, most of the linker mutants could form stable 1:1 complexes with human plasminogen.Taken together, these observations suggest that (i) the functioning of the inter-domain linkers of SK requiresa critical minimal length, (ii) linker 1 is relatively more tolerant to insertions and sequence alterations, andappears to function primarily as a covalent connector between the α and β domains, and (iii) the nativelinker 2 sequence is virtually indispensable for the activity of SK probably because of structural and/orflexibility requirements in SK action during catalysis.

an plasminogen; HPN, humanio-β-D-galactopyranoside; Kd,kon, rate of association; μPG,, p-nitrophenyl p-guanidino-A; SK, streptokinase; SOE-PCR,R, surface plasmon resonance;similis expressed in E. coliand Engineering, Institute ofIndia. Tel.: +91 172 2690785;

ll rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

Streptokinase (SK) is an extra-cellular protein secreted by severalspecies of the genus Streptococcus that has the capability of activatingthe human plasminogen (HPG). HPG is one of the key components ofthe fibrinolytic system that, once converted to plasmin (HPN) helps todissolve pathological blood clots [1,2]. Streptokinase consists of threestructurally homologous, independently folded domains (α, β and γ)connected by two flexible linkers [3,4]. SK functions principally as a

protein cofactor that forms a high-affinity equimolar complex withHPG (SK.HPG). Immediately upon the binding of SK conformationalchange/s is induced in ‘partner’HPG and opens its active site resultingin an enzymatically active complex (termed SK.HPG*), which thenrapidly converts to the mature form, SK.HPN [5,6]. The activatorcomplex has limited substrate specificity for “free” molecules ofplasminogen, converting them into plasmin [7,8].

Most studies on SK action to date have focused on identifying, andthen mechanistically exploring, the importance of critical epitopes inthis protein that are involved either in zymogen activation in thenascent SK.HPG complex [9], or ones involved in SK.HPG complexa-tion per se [10–12]. It has been observed that isolated SK domains (i.e.α, β or γ), to some extent, and its two-domain derivatives (αβ or βγ)retain significant capability to bind with HPG in a native-like manner,but have little HPG activator activity of their own, thereby suggestingthat the presence of all three contiguous domains is required for theefficient generation of this ‘higher order’ catalytic functionality in themolecule [8,13]. In solution, free SK has been shown to be highlyflexible, with the three domains being able to take many differentconformations presumably due to free movements in the linking

1731S. Yadav, G. Sahni / Biochimica et Biophysica Acta 1804 (2010) 1730–1737

sequences [14]. SK assumes a highly structured configuration only onbinding to its partner, either full-length HPG, or its isolated catalyticdomain μPG, which is devoid of the five kringle domains. Perhapsbecause of its inherent intramolecular flexibility, SK alone could neverbe crystallized, but only once it forms a bimolecular complexwith μPN[5]. Hence, any conformational change/s that might occur in SK interms of the relative orientation of its three domains after binarycomplex formation are not known compared to their preferredorientations – if indeed any preferred orientations exist – inuncomplexed SK in solution. In this context, there is a need for theavailability of mutants of SK that may be specifically affected in termsof critical conformational changes required for catalysis. Linkers inmulti-domain proteins are believed to control favorable and unfavor-able interactions between adjacent domains by means of the variableflexibility allowed by their primary structures [15,16]. The crystalstructure of SK.μPN [5], apart from revealing the linkers as two shortcoiled coils that connect the three domains of SK, does not, by itself,suggest any overt structure–function role of linkers.

Currently thus, there exists a gaping lack of even the mostrudimentary information on the role of the two inter-domain SKlinkers. A better understanding of the residues composing linkerswould allow us to introduce fluorescent labels in the linkers at rightpositions and thus would help us to follow the dynamics of thedomains of SK during the substrate binding or release. In order tobegin to investigate their role, we constructed various substitution,deletion and insertion mutants of the two linkers of SK, and analyzedthese for their functionality. The results, presented below, suggestthat linker 2, which connects β and γ domains, unlike linker 1 whichappears to function mainly as a flexible connector segment betweenthe alpha and beta domains, has a much stricter primary structurerequirement likely due to the presence of a relatively constrained,highly sequence-dependent conformation.

2. Materials and methods

2.1. Reagents

Glu-plasminogen was either purchased from Roche Diagnostics,GmbH (Penzberg, Germany) or purified from human plasma byaffinity chromatography [17]. pET-23d and Escherichia coli strain BL21(DE3) were products of Novagen Inc. (Madison, WI). ThermostableDNA polymerase (Pfu) was obtained from Stratagene Inc. (La Jolla,CA), and restriction endonucleases and other DNAmodifying enzymeswere purchased from New England Biolabs (Beverly, MA). Oligonu-cleotide primers were supplied by Biobasic, Inc., Canada. TheQuikChange™ Site-Directed Mutagenesis Kit was acquired fromStratagene, and used according to the manufacturer's instructions.Phenyl-Agarose 6XL was procured from Prometic Biosciences Ltd.,U.K. All other reagents used were of the highest analytical gradeavailable.

2.2. Cloning and mutagenesis

The linker substitution and deletion mutants were constructedusing mutagenesis experiments based on both splicing overlapextension-PCR (SOE-PCR) and QuikChange mutagenesis strategy.The mutants were cloned in pET23(d) vector in BL21 (DE3) hoststrain. All the linker mutant constructs were confirmed by DNAsequencing.

2.3. Protein expression and purification

The rSK-linker mutant constructs were expressed intra-cellularly,being derived from the SK gene of S. equisimilis strain H46A (Gene-bank accession no. gb/K02986/STRSKC). The procedures describingits cloning in E. coli BL21 (DE3) cells and expression as inclusion

bodies (IBs) under the control of the T7 phage RNA polymerasepromoter after induction with isopropyl-1-thio-β-D-galactopyrano-side (IPTG) are described earlier [18]. Briefly, the cells were pelletedafter 6–7 h of induction at 40 °C and were lysed through sonication.The pelleted and washed IBs were then dissolved in 8 M urea andprotein folded after a rapid 10-fold dilution. The sample was thenchromatographed on Phenyl-Agarose 6XL beads and eluted in water.The obtained fractions were further purified by ion-exchangechromatography on DEAE-Sepharose® Fast-Flow, wherein highlypurified SK, or its mutants, eluted in the 0.15–0.2 M range. All SKproteins eluted were generally ∼95% pure, as analyzed by SDS-PAGE.Protein sequencing revealed that 60–70% of the N-termini of the SK, orits mutants, expressed and purified as described above had Ile (the N-terminus of mature SK), presumably due to processing by endogenousmethionine amino peptidase.

2.4. Preparation of HPN

HPN, the active form of HPG, was prepared by digesting Glu-HPGwith UK covalently immobilized on agarose beads [11] using a ratio of300 Plough units/mg HPG in 50 mM Tris.Cl, pH 8.0, 25% (v/v) glyceroland 25 mM L-Lysine at 22 °C for 10 h [18].

2.5. Proteolytic stability of rSK/SK-linker mutants

Purified rSK- or the rSK-linker mutants were pre-incubated withequimolar HPN (5 μM each) at 22 °C in 50 mM Tris.Cl assay buffer, forvarying time periods (0, 2, 5, 8 and 10 min) to check for the stability ofSK/linker mutants [7]. Samples were withdrawn after each timepoint, and the reaction was stopped by addition of 10 mM NPGB (tostop HPN-mediated proteolysis) [19,20] followed by same volume of2× sample buffer. These were then analyzed for an extent ofproteolytic degradation on a 10% SDS-PAGE.

2.6. Determination of plasminogen activation activity

A one-stage assaymethodwas used tomeasure the kinetics of HPGactivation by rSK or its mutants [21,22]. Purified rSK or mutant SK(0.5–5 nM) was added to a 100 µl quartz assay cuvette containing1 µM HPG in an assay buffer (50 mM Tris.Cl buffer, pH 7.5) alsocontaining 1 mM chromogenic substrate (Chromozym® PL, Tosyl-glycyl-prolyl-lysine-4-nitranilide-acetate, Roche Diagnostics GmbH,Germany). The change in absorbance at 405 nmwas thenmeasured asa function of time (t) at 22 °C. The activator activities were obtainedfrom the slopes of the activation progress curves, which were plottedas change in absorbance/t against t [21].

2.7. Determination of kinetic constants for plasminogen activationactivity

The kinetics of HPG activation by HPN.SK/SK-linker mutantcomplexes were measured by transferring suitable aliquots ofpreformed mutant complexes with equimolar HPN to the assaycuvette containing different concentrations of substrate HPG [21] inassay buffer (50 mM Tris.Cl buffer, pH 7.5) also containing 0.5 mMchromogenic substrate (Chromozym® PL, Tosyl-glycyl-prolyl-lysine-4-nitranilide-acetate, Roche Diagnostics GmbH, Germany). Thegeneration of activator activity was monitored at 22 °C at 405 nm.The kinetic parameters for HPG activation were then calculated frominverse Lineweaver–Burk plots [21].

2.8. Esterolytic activation of equimolar HPG.SK/SK-loop complexes

To monitor the active site formation 7 μM HPG was added to anassay cuvette containing 7.5 μM rSK/SK mutant, 100 μM NPGB (p-nitrophenyl p-guanidinobenzoate) and 10 mM sodium phosphate

Table 1Substitution mutants of the two linkers of rSK.

Linker 1mutants

Mutation Linker 2mutants

Mutation

Wt-linker 1 α–YKEKP IQNQAKS–β Wt-linker 2 β–K K G E K P Y D P F D RS H L–γ

SKlin2–lin1 α–K K G E K P Y D P F D RS H L–β

SKlin2–lin1 β–Y K EK P I QNQAK S–γ

SKlin2–lin2 Y K E K P I Q N Q A K S SKlin1–lin1 KK GE KP YD PF DR SH LKK GE KP YD PF DR SH L Y K E K P I Q N Q A K S

SKlin1(Gly)9 Y K E K P I Q N Q A K S SKlin2(Gly)9 KK GE KP YD PF DR SH LG G G G G G G G G G G G G G G G G G

SKlin1–QNQ–AAA Y K E K P I Q N Q A K S SKlin2–YFS–AAA KKGEKPYDP FDR SHLA A A A A A

Columns 1 and 2 represent the substitution mutants of linker 1, and the correspondinglinker 2 mutants are represented in columns 3 and 4. The wild type sequence of thelinker substituted is shown at the top in bold, and the residues substituted arementioned below. The residues added to the native linker sequence or in the non-native linker sequence are marked in bold.

Table 2Plasminogen activation activity by complexes of HPG.rSK/rSK-linker swapped andsubstitution derivatives.

Linker 1mutants

% HPG activationactivity

Linker 2mutants

% HPG activationactivity

Wt-SK 100 Wt-SK 100SKlin2–lin1 0.5 SKlin2–lin1 0.5SKlin2–lin2 80 SKlin1–lin1 11SKlin1(Gly)9 2 SKlin2(Gly)9 n.o.SKlin1–QNQ–AAA 90 SKlin2–YFS–AAA 20

“n.o.” denotes that data could not be obtained due to absence of any activity.

1732 S. Yadav, G. Sahni / Biochimica et Biophysica Acta 1804 (2010) 1730–1737

buffer, pH 7.5. The ‘burst’ of p-nitrophenol release due to acylation ofan active center was monitored at 410 nm as a function of time at22 °C [19,23]. Purified streptokinase, or the various mutants tested foresterolytic capability gave characteristic NPGB ‘bursts’ due to N-terminal methionine processing (see above), and when treated invitro with pfu derived methionine amino peptidase (data not shown)there was only a very marginal increase in the maximal levels of theesterolytic bursts with NPGB.

2.9. Kinetic analysis of protein–protein interactions using surfaceplasmon resonance (SPR)

The rates of association and dissociation between HPG and therSK/SK-linker mutants were followed in real time by SPR using theIAsys Plus™ system (Cambridge, UK) [8,24–26]. In these experiments,the carboxy methyl-dextran (CMD) cuvette was activated usingamino-coupling EDC-NHS chemistry according to the manufacturer'sprotocols (IAsys protocol 1.1). This was followed by the immobiliza-tion of HPG to the activated CMD cuvette. Non-specifically bound HPGwas then removed by repeated washings with 10 mM phosphatebuffered saline (PBS), pH 7.4, followed by three pulses given with10 mM HCl. The net response chosen for the immobilized HPG ontothe cuvette was around 800 arc sec in all experiments. Experimentswere performed at 22 °C in PBS, pH 7.4 binding buffer containing0.05% Tween 20 and 50 mMNPGB. The latter was included in order toprevent HPN-mediated proteolysis [20].

After equilibrating the cuvette with binding buffer, varyingconcentrations of either rSK or rSK-linker mutants were added andeach binding response was monitored during the ‘association’ phase.Subsequently, the cuvette was washed with binding buffer and the‘dissociation’ phase was recorded [27]. Following each cycle ofanalysis, the cuvette was regenerated by washing with 10 mM HCl,and baseline was re-established with binding buffer. In parallel, in thecontrol cell of the dual channel cuvette, activated CMD alone wastaken as a negative control for the binding studies.

The data were analyzed after subtraction of the correspondingnon-specific refractive index component(s), and the kinetic constantswere calculated from the sensorgrams by non-linear fittings ofassociation and dissociation curves using the software FASTfit™,supplied by the manufacturer. The equilibrium dissociation constant(Kd) was then calculated from the extent of association of monophasiccurve. The kon was calculated from the equation koff/Kd [28].

3. Results

3.1. Effect of swapping the inter-domain linker sequences

The SK linkers show overall homology and span an approximatelyequal spatial distance between their respective domains (Fig. 1). Itwas of interest to examine if they could be exchanged amongstthemselves and still be able to sustain a functional SK molecule. Usingthe overlap extension strategy (seeMaterials andmethods), the linkersequences were interchanged i.e. linker 1 placed at linker 2 position,namely between β and γ domains, and linker 2 placed betweenα andβ domains, resulting in the construct SKlin2–lin1 (Table 1). The mutantSKlin2–lin1 was then tested for capability to activate substrate HPG [21].The assay showed that the “cross-substituted” mutant was almost200-fold less active than nSK (Table 2). The possible reason could bethat the swapping of linkers created/exposed new proteolyticcleavage sites in the resultant molecule, so that during the assay themutant underwent rapid degradation and consequent loss in activity.Hence, the linker-swapped mutant was checked for its stability inequimolar complex with HPN (see Materials and methods), and thepattern of cleavage was observed by SDS-PAGE. The earlier studieshave revealed that the incubation of nSK with PG/PN results in aspecific pattern of degradation and suggest that two fragments are

rapidly cleaved from SK, from the N-terminus (residues 1–59) and theC-terminus (residues 385–414) [29]. The resultant SK fragment of a37 kDa band corresponding to SK 60–385, marked as SK-h in Fig. 1 ofthe paper, is capable of substrate activation as it is capable of bindingboth partner plasmin(ogen) and its complementary fragment, SK 1–59 [22,30]. However, unlike nSK, the 37 kDa band was not observedfor SKlin2–lin1, thus suggesting that the linker-swappedmutant SKlin2–-lin1 was highly susceptible to PN mediated digestion, leading to therapid generation of smaller degraded fragments (Fig. 2 panel B). Theproteolytically unstable mutant was not characterized any further.Nevertheless, it is evident that the two SK linkers are not directlyinterchangeable despite an apparent sequence homology betweenthe two.

3.2. Substitution of the two linkers with native and non-native sequences

In view of the sequence homologies that the two linkers share, itneeded to be explored whether one could be preferentially put intothe other's place, even if a direct cross-substitution was untenable. Totest this experimentally, linker 1 sequencewas substitutedwith linker2 sequence, resulting in a rSK mutant with two linker 1 sequences,namely SKlin1–lin1. Similarly, another SK gene with two linker 2sequences, namely SKlin2–lin2 was also constructed (Table 1). Simul-taneously, both the linker sequences were also substituted (inseparate SK constructs) with poly-(glycine) nona-peptide sequences,resulting in constructs SKlin1(Gly)9 and SKlin2(Gly)9 (Table 1). The lengthof poly-Gly polypeptide segment was selected on the basis of thedistance that is seen to be spanned by the two inter-domain linkers inthe SK.μPN crystal structure, which is nearly 25 Å. The poly-Glysequence is known to confer an extended coil conformation [31], andthus could be expected to substitute the two SK linkers without

Fig. 1. Ribbon diagram of SK depicting the two inter-domain linkers. A ribbon diagramof SK derived from the crystal structure SK.μPN (PDB — 1BML), showing the threedomains of SK α, β, and γ linked by the two flexible coiled coil linkers, namely linker 1and linker 2. The figure (inset) also shows the close homology between the sequencesof the two linkers, which highlights the presence of several charged and hydrophilicresidues in both linkers. The ribbon diagram was made using PyMol software andshows the almost equal spatial distance (25 Å) spanned by the two linkers in the X-raystructure.

1733S. Yadav, G. Sahni / Biochimica et Biophysica Acta 1804 (2010) 1730–1737

disturbing their native conformation if only flexibility characteristicswere sufficient to ensure linker functioning.

All the substitution mutants were constructed using the overlapextension strategy, as before, and the purified proteins were checkedfor their HPG activation activities (summarized in Table 2). The resultsshow that the substitution mutant SKlin2–lin2 exhibited almost WT-rSK-like activity, suggesting that linker 1 can indeed be substituted bylinker 2. In contrast, the SKlin1–lin1 mutant showed a 10-fold decreasein HPG activation capability. The mutants were then furthercharacterized by kinetic studies with substrate PG (summarized inTable 3). The data show that the substitution mutant SKlin2–lin2

exhibited almost wild type-like cofactor ability. In contrast, theSKlin1–lin1 mutant showed a 10-fold decrease in substrate activationcapability and a 4-fold increase in km values for HPG as substrate(Table 3), suggesting that linker 2 cannot be substituted by linker 1sequence as that affect the catalytic efficiency of the resultantconstruct. The substitution mutants were then studied for theirproteolytic stability as described before, wherein both SKlin1–lin1 andSKlin2–lin2 mutants showed WT-rSK-like stability (Fig. 2). Thus, thedecreased cofactor activity of SKlin1–lin1 could be due to either a lack ofaffinity between the mutant and HPG to form a suitably tight activebimolecular complex, or could be due to an alteration at the substrateprocessivity step. In order to distinguish between these twopossibilities, the linker mutants were first studied for their ability tomake an active binary complex as indicated by their ability to openthe active site in partner HPG (via the so-called Pathway I) using theactive site acylating agent, NPGB. NPGB has been previously used totitrate the active site formation in the SK.HPG complex [19,23].Neither SK nor HPG alone gives the ‘burst’ observed at 410 nm, but

Table 3Steady-state kinetic parameters for HPG activation by equimolar complexes of HPN.rSK/rSK-linker swapped and substitution derivatives.

Linker 1mutants

kmμM

kcatmin−1

Linker 2mutants

kmμM

kcatmin−1

Wt-SK 0.5±0.05 11±0.5 Wt-SK 0.5±0.05 11±0.5SKlin2–lin2 0.8±0.1 11±0.05 SKlin1–lin1 2±0.05 1.5±0.01SKlin1(Gly)9 0.7±0.05 0.11±0.05 SKlin2(Gly)9 n.d.SKlin1–QNQ–AAA 0.5±0.05 9.35±0.02 SKlin2–YFS–AAA 0.8±0.2 2.2±0.03

“n.d.” denotes not determined.

only when the two are mixed in equimolar proportions that they gavea characteristic burst. This burst is associated with rapid NPGBhydrolysis, observed due to the formation of a ‘virgin’ SK.HPGactivator complex and consequent acylation of the active site.Interestingly, the SK-linker substitution mutants behaved in a closelyWT-like mode (Fig. 3), suggesting that the SKlin1–lin1 mutant couldform a stable and active binary complex with the partner molecule.

To further quantitatively measure the binary binding step, a directphysico-chemical approach, namely Surface Plasmon Resonance(SPR) was adopted [24,25]. These studies were done in the presenceof the inhibitor, NPGB, by procedures earlier established [13,20,26,32]to prevent the enzymatic activation of the nascent complex during theSPR experiments. The data on the interaction of rSK, or its mutants,with HPG were analyzed and the kinetic constants were calculatedfrom the sensorgrams by non-linear fittings of association anddissociation curves [8,26] using the software FASTfit™, supplied bythe manufacturer. The association–dissociation parameters soobtained for the binary interactions (Table 4) clearly showed thatthe affinity of the SKlin1–lin1 construct was closely similar to that of rSKfor HPG. These results essentially suggest that the binding of the linkersubstitution mutant with HPG was not perturbed.

The results of plasminogen activation activity measurements for thepoly-gly substitutionmutant for linker 1, SKlin1(Gly)9, showed nearly 50-fold lower activity compared to nSK, while the SKlin2(Gly)9 showedalmost no activity (Table 2). Similar results were observed from thekinetic study of these mutants (Table 3). The poly-Gly substitutionmutantswhen checked for their proteolytic stability, showed thatunlikelinker 1, the linker 2 substitution (SKlin2(Gly)9) showed deleteriouseffects (Fig. 2). The proteolytic susceptibility of the poly-Gly substitutionmutant could be predicted to be due to conformational changes in themutant construct since the poly-Gly sequence per se did not introduceany new plasmin-sensitive residues in the linker region.

3.3. Study of linker length by deletion and linker insertion mutagenesis

To address the possibility of residue number being one of thecausative factors of activity loss with the substitution mutants, theeffect of systematic deletion of the length of the linkers was thenexamined (see Table 5). The deletion of a single residue from themiddle of linker 1 resulted in no measurable change in activity forsubstrate PG activation, whereas for linker 2, a measurable decrease inactivity was observed even after a single-residue deletion (Table 6and 7). Deletion of three consecutive residues in linker 1 resulted inabout 3-fold decrease in activity, while in case of linker 2 about 5-folddecrease in HPG activation activity was observed. However, deletionof five residues from either of the linkers resulted in completeinactivation of the molecule (Table 6 and 7). The proteolytic stabilityanalysis of the deletion mutants showed that deletion beyond threeresidues in either of the linkers resulted in enhanced proteolyticsusceptibility. The unstable mutants were not characterized anyfurther; however, the mutants that retained stability and yet showeda decreased cofactor activity were studied further for their ability toform an active binary complex. The results (Fig. 3 and Table 4) suggestthat deletions of up to 3 residues from either linker neither perturbthe resultant mutants' active site induction in partner HPG nor dothese affect their 1:1 binary binding constants with HPG.

In general, comparatively more severe effects were observed forthe linker 2 deletion constructs compared to those of linker 1. In orderto explore whether an increase in their length was detrimental or not,we inserted short non-native sequences so as to promote eitherflexible or relatively rigid local structural conformations into thelinker sequences. The pentapeptide sequence, SAESG, is known toadopt an extended conformation [33], whereas oligomeric prolinesequences [e.g. (Pro)5] are known to favor a rigid polyproline II helix[34]. The SAESG sequence was selected from the database ‘LINKER’that describes peptide sequences that are known to adopt extended

Fig. 2. Proteolytic stability of rSK/rSK-linker mutants. Relative proteolytic stability of rSK/SK-linker mutants in the presence of stoichiometric levels of HPN as detected by SDS-PAGE.Samples were withdrawn periodically from equimolar mixtures of each protein with HPN, quenched with inhibitor NPGB, and the proteolysis of SK and SK-linker mutants atdifferent time-intervals (0, 2, 5, 8 and 10 min after complexation) as shown by lanes in the four panels (A, B, C and D as indicated), was analyzed on SDS-PAGE, as detailed underMaterials and methods. Lanes marked c, in each case, represent the control reactions in which the proteins were mixed only after addition of NPGB showing the HPN and SK/SK-linker mutant bands at their respective positions, as also indicated bymolecular weight markers, run alongside in parallel lanes. Bands corresponding to the HPN heavy chain (PN-h),light chain (PN-l), native SK, and the degradation fragments of SK (SK-d) are also indicated. Panel A shows the profile of rSK, while panel B shows the unstable mutant SKlin2–lin1.Panels C and D show the stable mutants SKlin1(Gly)9 and SKlin2–YFS–AAA, with profiles similar to that of WT-rSK.

1734 S. Yadav, G. Sahni / Biochimica et Biophysica Acta 1804 (2010) 1730–1737

conformations as confirmed by X-ray crystallography and NMR [33].The SAESG pentapeptide is a part of the linker region of phenolhydroxylase (1FOH) of Trichosporon cutaneum and is composed offlexible, linker-favoring residues.

As shown in Table 5, the insertion of the two short sequences wasmade at two different positions, i.e. either at the N- or the C-terminalregion of the two SK linkers. Both SAESG and the (Pro)5 sequences wereinserted in linker 1 between the residues E and K (to obtain the mutantstermedasSKlin1+SAESG(E/K) andSKlin1+PPPPP(E/K), respectively), orbetweenresidues Q and A in the same linker (to obtain mutants SKlin1+SAESG(Q/A)

and SKlin1+PPPPP(Q/A)) as described in Table 5. Similarly, these twoextraneous sequenceswere incorporated into linker 2between residues Eand K (to obtain mutants SKlin2+SAESG(E/K) and SKlin2+PPPPP(E/K)) orbetween residuesRandS(mutants SKlin2+SAESG(R/S) andSKlin2+PPPPP(R/S)).

The results of the plasminogen activation and the kinetic assayscarried out with these mutants (Table 6 and 7) reveal that the linker 1

Table 4Binary interaction of HPG with rSK/rSK-linker mutants as analytes.

Linker 1 mutants kon(×106) (M−1 s−1)

koff(×10−2) (s−1)

Kd

(×10−9) (M)

Wt-SK 1.4±0.3 2±0.2 14±3SKlin2–lin2 1.2±0.2 2.3±0.2 19±4SKlin1(Gly)9 1.2±0.1 1.9±0.1 16±1SKlin1–N1del 1.3±0.2 2.1±0.3 16±3SKlin1–N3del 1.5±0.2 2.7±0.4 17±2SKlin1+SAESG(E/K) 1.3±0.2 2±0.3 15±3SKlin1+PPPPP-(E/K) 1.4±0.2 2±0.1 14±4SKlin1–QNQ–AAA 1.3±0.1 2±0.1 15±1

Equilibrium dissociation constants of interaction between HPG immobilized on CMD cuvette“n.d.” denotes not determined.

insertion mutants did not undergo any major loss in activitycompared to WT-rSK, neither with the SAESG insert nor with the(Pro)5 insert, suggesting that linker 1 is tolerant to the insertion ofboth an ‘extending’ sequence, as well as a conformationallyconstraining oligo-Pro sequence. The kcat values with linker 2insertions suggest that it is relatively tolerant to an increase in itslength due to insertion of the SAESG sequence (Tables 6 and 7).However, an increase in the linker 2 length due to the insertion of theoligo-Pro sequence has unfavorable consequences on the functioningof the resultant SKmolecule. It can be seen that the insertion of penta-Pro, between E and K residues (SKlin2+PPPPP(E/K)), decreased theactivity by almost 10-fold, but its insertion at the C-terminal region,between R and S residues (SKlin2+PPPPP(R/S)), resulted in a much moredrastic decrease in activity (almost 100-fold) (Table 6 and 7). Theinsertions in linker 2 also slightly increased the km values forsubstrate. These results suggest that extraneous sequence insertion

Linker 2 mutants kon(×106) (M−1 s−1)

koff(×10−2) (s−1)

Kd

(×10−9) (M)

Wt-SK 1.4±0.3 2±0.2 14±3SKlin1–lin1 1±0.2 2±0.1 20±5SKlin2(Gly)9 n.dSKlin2–N1del 1.5±0.1 2.7±0.3 18±2SKlin2–N3del 1.2±0.2 2±0.2 17±3SKlin2+SAESG(E/K) 1.3±0.2 2.5±0.1 19±4SKlin2+PPPPP-(E/K) 1.6±0.2 2±0.12 12±4SKlin2–YFS–AAA 1±0.1 2±0.2 20±7

with SK/SK-linker mutants as binary analyte at varied concentrations were determined.

Table 5The deletion and insertion mutants of the rSK inter-domain linkers.

Linker 1 mutants Mutation Linker 2 mutants Mutation

Wt-linker 1 α–Y K E K P I Q N Q A K S–β Wt-linker 2 β–K K G E K P Y D P F D R S H L–γSKlin1–N1del Y K E K P I _N Q A K S SKlin2–N1del K K G E K P Y_P F D R S H LSKlin1–N3del Y K E K ___N Q A K SKlin2–N3del K K G E K ___P F D R S H LSKlin1–N5del Y K _____N Q A K SKlin2–N5del K K G _____P F D R S H LSKlin1+SAESG(E/K) Y K E-SAESG-K P I Q N Q A K S SKlin2+SAESG(E/K) K K G E-SAESG-K P Y D P F D R S HLSKlin1+SAESG(Q/A) Y K E K P I Q N Q-SAESG-A K S SKlin2+SAESG(R/S) K K G E K P Y D P F D R-SAESG-S HLSKlin1+PPPPP(E/K) Y K E-PPPPP-K P I Q N Q A K S SKlin2+PPPPP(E/K) K K G E-PPPPP-K P Y D P F D R S H LSKlin1+PPPPP(Q/A) Y K E K P I Q N Q-PPPPP-A K S SKlin2+PPPPP(R/S) K K G E K P Y D P F D R-PPPPP-S H L

The table shows, in columns 1 and 2, the linker 1 mutants, whereas columns 3 and 4 list the linker 2 mutants. Deletion of residues from the respective native linkers has beenrepresented as gaps between the residues flanking a given deletion. The insertions into the native sequences of both the linkers are marked in bold type.

1735S. Yadav, G. Sahni / Biochimica et Biophysica Acta 1804 (2010) 1730–1737

in linker 2 is position-specific, and any torsional constraint caused byinserting a rigid linker sequence, like (pro)5, results in virtuallycomplete loss in activity.

In order to analyze the catalytic step that is compromised in theinsertion mutants, their proteolytic susceptibility was checked, asbefore, in the presence of HPN. The reactions were analyzed by SDS-PAGE (Fig. 2) which show that all the linker insertion mutantsexhibited WT-rSK-like stability except the mutant SKlin2+PPPPP(R/S),which showed complete degradation in the presence of HPN underconditions where the wild type is relatively stable. These resultsstrongly suggest that the functioning of linker 2 seems to be muchmore conformation-dependent. The linker insertion mutants exceptSKlin2+PPPPP(R/S), were then further studied for their ability to formactive binary complexes. The results clearly show that the mutantshad a native-like partner HPG activation (Fig. 3) and binding affinity(Table 4).

3.4. Study of the role of critical residues constituting the two linkers

Both the inter-domain linkers of SK are composed of severalcharged and hydrophilic residues that are solvent exposed asobserved in the SK.μPN crystal structure coordinates [5]. We firstinvestigated the possible involvement of the charged residues bysubstitution with alanines followed by analysis for substrate activa-tion capability. However, none of the charged substitutions of eitherlinkers showed any significant change in activity (data not shown),suggesting that the linkers might not be involved in any charge-mediated interactions that are catalytically important.

To further investigate the role of the side chains of linkers thatmight be involved in some other kind of non-covalent interaction byvirtue of their high solvent accessibility, several native residues weresubstituted with alanine residues. The mutant SKlin1–QNQ–AAA showedactivity that was native-like (Table 1); however, mutant SKlin2–YFS–AAA

showed a 5-fold drop in the activity (Table 2). The structure of linker2, as evident from the SK.μPN crystal structure, shows that these threeresidues are involved in generating a kink in an otherwise outwardlyexposed linker, that faces the substrate binding pocket of the activator

Table 6Plasminogen activation activity by complexes of HPG-rSK/rSK-linker deletion andinsertion mutants.

Linker 1mutants

% HPG activationactivity

Linker 2mutants

% HPG activationactivity

Wt-SK 100 Wt-SK 100SKlin1–N1del 100 SKlin2–N1del 60SKlin1–N3del 30 SKlin2–N3del 20SKlin1–N5del n.o SKlin2–N5del n.o.SKlin1+SAESG(E/K) 100 SKlin2+SAESG(E/K) 85SKlin1+SAESG(Q/A) 100 SKlin2+SAESG(R/S) 80SKlin1+PPPPP(E/K) 85 SKlin2+PPPPP(E/K) 7SKlin1+PPPPP(Q/A) 90 SKlin2+PPPPP(R/S) n.o.

“n.o.” denotes that data could not be obtained due to absence of any detectable activity.

complex. The single and double mutants of YFS residues were thenconstructed to identify the most critical residue/s out of the three.However, none of these mutants showed any significant change inactivity in terms of substrate activation capability compared to wildtype rSK (data not shown), suggesting that only when all threeresidues are mutated together does a major alteration in cofactoractivity ensues. These results are supportive of the earlier conclusionsfrom insertion and substitution experiments on the linkers' primarystructure requirements that unlike linker 1, linker 2 likely requires adefined conformation and is not merely a randomly coiled and/orflexible linker connecting two SK domains.

4. Discussion

The crystal structure of SK [5] suggests that the SK linkers aresurface-accessible and have a high coiled coil propensity. Both thelinkers show sequence homology with each other and are rich inhydrophilic and charged residues (Fig. 1). However when the linkerswere swapped amongst themselves the resultant molecule lost itsstability, likely resulting in its drastically low cofactor activity. Hence,to further investigate the role of linkers in SK functioning, variouslinker substitution mutants were constructed. The SK construct inwhich linker 2 was substituted with linker 1 sequence resulted in alow activity molecule with a decreased affinity for substrate. This issuggestive of linker 2 being involved in substrate interactions eitherdirectly, or indirectly, by changing the orientation of the flankingdomains (without affecting 1:1 binary complexation). However,interestingly the SK molecule with two linker-2 sequences behavedlike WT-rSK. Further substitution studies of linkers with poly-glycineresidues also suggest that the substitution of linker 1 resulted indrastic reduction in activity although substitution of linker 2 resultedin a proteolytically labile mutant. This study amply illustrates thatlinker 1, holding together the α and β domains of SK, is not highlysequence-specific. However, linker 2, serving as a connector betweenthe β and γ domains, exhibits a pronounced sequence-dependence.

The linker mutants that showed altered plasminogen activationactivities but were proteolytically stable were further checked fortheir ability to form binary complex with plasminogen through thewell known NPGB assay. The NPGB ‘burst’ observed with the linkermutants clearly suggests that these mutants are capable of forming anactive binary complex with partner HPG. The direct binary studiesdone using SPR further confirm the conclusion that although thesubstitution of linkers does affect the activity of the molecule this isnot due to any effect on the binary complex formation. This suggeststhat it is mainly the structural properties of the three domains of SK,rather than any of the linker side chains (once a certain ‘minimal’length is preserved), that mainly contribute towards the generation ofthe 1:1 complex with partner HPG. However, once the complex isformed, the substrate processivity, that follows this primary bindingevent between SK and HPG, seems to be adversely affected bysubstitution of the linker 2 by linker 1, resulting in a decreasedcofactor activity against substrate HPG.

Table 7Steady-state kinetic parameters for HPG activation by equimolar complexes of HPN-rSK/rSK-linker deletion and insertion mutants.

Linker 1 mutants km(μM)

kcat(min−1)

Linker 2 mutants km(μM)

kcat(min−1)

Wt-SK 0.5±0.05 11±0.5 Wt-SK 0.5±0.05 11±0.5SKlin1–N1del 0.43±0.04 9.9±0.01 SKlin2–N1del 0.66±0.03 6.6±0.04SKlin1–N3del 0.6±0.02 2.75±0.005 SKlin2–N3del 1±0.2 0.99±0.01SKlin1+SAESG(E/K) 0.45±0.05 10.45±0.01 SKlin2+SAESG(E/K) 1±0.2 4.4±0.02SKlin1+SAESG(Q/A) 0.5±0.1 8.8±0.03 SKlin2+SAESG(R/S) 0.53±0.05 4.95±0.03SKlin1+PPPPP(E/K) 0.4±0.05 8.25±0.03 SKlin2+PPPPP(E/K) 1±0.2 0.55±0.05SKlin1+PPPPP(Q/A) 0.5±0.03 9.35±0.02 SKlin2+PPPPP(R/S) n.d

“n.d.” denotes not determined.

1736 S. Yadav, G. Sahni / Biochimica et Biophysica Acta 1804 (2010) 1730–1737

The experiments described in the foregoing sections stronglysuggest that linker 1 functioning is not critically dependent on itsnative primary structure, and can undergo a variety of alterationsprovided that the linker length is not being compromised beyond aminimal value. However, linker 2, unlike linker 1, seems to be moresequence-specific and does not tolerate substitution by a wide varietyof sequences that were quite well-tolerated at the linker 1 position.Hence, it was of interest to explore the role of key residues in thislinker through site-specific substitution mutagenesis. The alanine-substitution of charged and exposed residues of linker 1 behaved likeWT-rSK. However, the linker 2 substitution, of YFS residues, resultedin decreased cofactor activity as suggested by the steady-state kineticdata. In summation, it appears that compared to linker 2, the role oflinker 1 of SK is more restricted, in that it acts principally like acovalent “stuffer” that servesmainly to hold the two flanking domainsat an optimal distance and with sufficient flexibility.

One of the main insights that emerge from this study is thatsequences of both linkers (but more so in case of linker 2) are at leastindirectly involved in enzyme–substrate interactions. As many of thefunctionally attenuated linker mutants were found to be mechanis-tically deficient at the enzyme–substrate interaction stage, butexhibiting WT-like characteristics in terms of their capabilities forthe primary step of HPG.SK binary complexation. These events (atpresent largely speculative in the absence of direct experimentalevidence) could be the intra- and inter-molecular and/or inter-domain movements facilitating substrate processivity. This study alsohelps in identifying the ‘allowed’ and ‘unallowed’ residues in the

Fig. 3. Esterolytic activation of equimolar HPG.rSK/rSK-linker complexes. Active sitetitration of HPG on complexing with nSK and SK-linker mutants using NPGB. Thegeneration of an active center in “partner” HPG by rSK, and linker mutants, was moni-tored for release of a burst of p-nitrophenol as described under Materials and methods.The figure depicts the progress curves of NPGB hydrolysis by rSK-HPG (squares),SKlin1–lin1-HPG (up-triangles), SKlin1(Gly)9-HPG (down-triangles), SKlin1–N1del-HPG(open squares), SKlin1+SAESG(E/K)-HPG (open circles), SKlin2–YFS–AAA-HPG (crosses)complexes and a control reaction (circles) fromwhich SKwas omitted while containingall the other reaction components.

linkers. This vital information gives us the capability in future studiesto insert fluorescent and/or other labels in the linkers at positions thatwould not affect the overall activity of the SK molecule. This wouldallow us to follow the dynamics of the reaction across the variouscatalytic stages such as substrate capture, target bond scission andproduct release. In other words, with such strategically labeled linkersof SK we can study the orientation changes that occur in SK once thepartner PG binds to it after making the initial enzymatic complex andalso during further stages of catalysis. Indeed, such biophysicalstudies, now made feasible with the availability of the various linkermutants of SK constructed and characterized in this study, arecurrently in progress to explore the molecular details of SKfunctioning.

Acknowledgements

The authors thank Ms Paramjit Kaur for her expert technicalassistance in protein sequencing and other general protein purifica-tion protocols, and the DNA Sequencing Core facility of IMTECH. Thiswork was supported from intra-mural grants from the Council ofScientific and Industrial Research, India.

References

[1] F.J. Castellino, Recent advances in the chemistry of the fibrinolytic system, Chem.Rev. 81 (1981) 431–446.

[2] M.A. Parry, X.C. Zhang, I. Bode, Molecular mechanisms of plasminogen activation:bacterial cofactors provide clues, Trend Biochem. Sci. 25 (2000) 53–59.

[3] F. Conejero-Lara, J. Parrado, A.I. Azuaga, R.A. Smith, C.P. Ponting, C.M.Dobson, Thermal stability of the three domains of streptokinase studied bycircular dichroism and nuclear magnetic resonance, Protein Sci. 5 (1996)2583–2591.

[4] J. Parrado, F. Conejero-Lara, R.A. Smith, J.M. Marshall, C.P. Ponting, C.M. Dobson,The domain organization of streptokinase: nuclear magnetic resonance, circulardichroism, and functional characterization of proteolytic fragments, Protein Sci. 5(1996) 693–704.

[5] X. Wang, X. Lin, J.A. Loy, J. Tang, X.C. Zhang, Crystal structure of the catalyticdomain of human plasmin complexed with streptokinase, Science 281 (1998)1662–1665.

[6] G. Markus, W.C. Werkheiser, The interaction of streptokinase with plasminogen.Functional properties of the activated enzyme, J. Biol. Chem. 239 (1964)2637–2643.

[7] P.D. Boxrud, W.P. Fay, P.E. Bock, Streptokinase binds to human plasmin withhigh affinity, perturbs the plasmin active site, and induces expression ofa substrate recognition exosite for plasminogen, J. Biol. Chem. 275 (2000)14579–14589.

[8] V. Sundram, J.S. Nanda, K. Rajagopal, J. Dhar, A. Chaudhary, G. Sanhi, Domaintruncation studies reveal that the streptokinase–plasmin activator complexutilizes long range protein–protein interactions withmacromolecular substrate tomaximize catalytic turnover, J. Biol. Chem. 278 (2003) 30569–30577.

[9] X. Wang, S. Terzyan, J. Tang, J.A. Loy, X. Lin, X.C. Zhang, Human plasminogencatalytic domain undergoes an unusual conformational change upon activation,J. Mol. Biol. 295 (2000) 903–914.

[10] D. Nihalani, G. Sahni, Streptokinase contains two independent plasminogen-binding sites, Biochem. Biophys. Res. Commun. 217 (1995) 1245–1254.

[11] D. Nihalani, G.P. Raghava, G. Sahni, Mapping of the plasminogen binding site ofstreptokinase with short synthetic peptides, Protein Sci. 6 (1997) 1284–1292.

[12] K.C. Young, G.Y. Shi, Y.F. Chang, B.I. Chang, L.C. Chang, M.D. Lai, W.J. Chuang, H.L.Wu, Interaction of streptokinase and plasminogen. Studied with truncatedstreptokinase peptides, J. Biol. Chem. 270 (1995) 29601–29606.

1737S. Yadav, G. Sahni / Biochimica et Biophysica Acta 1804 (2010) 1730–1737

[13] J.A. Loy, X. Lin, M. Chenone, F.J. Castellino, X.C. Zhang, J. Tang, Domain interactionsbetween streptokinase and human plasminogen, Biochemistry 40 (2001)14686–14695.

[14] G. Damaschun,H. Damaschun, K. Gast, D. Gerlach, R.Misselwitz, H.Welfle, D. Zirwer,Streptokinase is a flexiblemulti-domain protein, Eur. Biophys. J. 20 (1992) 355–361.

[15] R.S. Gokhale, C. Khosla, Role of linkers in communication between proteinmodules, Curr. Opin. Chem. Biol. 4 (2000) 22–27.

[16] W.Wriggers, S. Chakravarty, P.A. Jennings, Control of protein functional dynamicsby peptide linkers, Biopolymers 80 (2005) 736–746.

[17] D.G. Deutsch, E.T. Mertz, Plasminogen: purification from human plasma by affinitychromatography, Science 170 (1970) 1095–1096.

[18] A. Chaudhary, S. Vasudha, K. Rajagopal, S.S. Komath, N. Garg, M. Yadav, S. Mande,G. Sahni, Function of the central domain of streptokinase in substrateplasminogen docking and processing revealed by site-directed mutagenesis,Protein Sci. 8 (1999) 2791–2805.

[19] T. Chase Jr, E. Shaw, Comparison of the esterase activities of trypsin, plasmin, andthrombin on guanidinobenzoate esters, titration of the enzymes, Biochemistry8 (1969) 2212–2224.

[20] F. Conejero-Lara, J. Parrado, A.I. Azuaga, C.M. Dobson, C.P. Ponting, Analysis of theinteractions between streptokinase domains and human plasminogen, ProteinSci. 7 (1998) 2190–2199.

[21] R.C. Wohl, L. Summaria, K.C. Robbins, Kinetics of activation of human plasminogenby different activator species at pH 7.4 and 37 degrees C, J. Biol. Chem. 255 (1980)2005–2013.

[22] G.Y. Shi, B.I. Chang, S.M. Chen, D.H. Wu, H.L. Wu, Function of streptokinasefragments in plasminogen activation, Biochem. J. 304 (1994) 235–241.

[23] R.C. Wohl, Interference of active site specific reagents in plasminogen-strepto-kinase active site formation, Biochemistry 23 (1984) 3799–3804.

[24] R. Cush, J.M. Cronin, W.J. Stewart, C.H. Maule, J. Molloy, N.J. Goddard, The resonantmirror: a novel optical biosensor for direct sensing of biomolecular interactionsPart I: Principle of operation and associated instrumentation, Biosens. Bioelectron.8 (1993) 347–354.

[25] P.E. Buckle, R.J. Davies, T. Kinning, D. Yeung, P.R. Edwards, D. Pollard-Knight, C.R.Lowe, The resonant mirror: a novel optical biosensor for direct sensing ofbiomolecular interactions Part II: applications, Biosens. Bioelectron. 8 (1993)365–370.

[26] J. Dhar, A.H. Pande, V. Sundram, J.S. Nanda, S.C. Mande, G. Sahni, Involvement of anine-residue loop of streptokinase in the generation of macromolecular substratespecificity by the activator complex through interaction with substrate kringledomains, J. Biol. Chem. 277 (2002) 13257–13267.

[27] D.G. Myszka, Kinetic analysis of macromolecular interactions using surfaceplasmon resonance biosensors, Curr. Opin. Biotechnol. 8 (1997) 50–57.

[28] T.A. Morton, D.G. Myszka, I.M. Chaiken, Interpreting complex binding kineticsfrom optical biosensors: a comparison of analysis by linearization, the inte-grated rate equation, and numerical integration, Anal. Biochem. 227 (1995)176–185.

[29] L. Summaria, L. Arzadon, P. Bernabe, K.C. Robbins, The interaction of streptokinasewith human, cat, dog, and rabbit plasminogens. The fragmentation of streptoki-nase in the equimolar plasminogen-streptokinase complexes, J. Biol. Chem. 249(1974) 4760–4769.

[30] F.J. Castellino, J.M. Sodetz, W.J. Brockway, G.E. Siefring Jr, Streptokinase, MethodsEnzymol. 45 (1976) 244–257.

[31] S. Ohnishi, H. Kamikubo, M. Onitsuka, M. Kataoka, D. Shortle, Conformationalpreference of polyglycine in solution to elongated structure, J. Am. Chem. Soc. 128(2006) 16338–16344.

[32] R.C. Wohl, L. Arzadon, L. Summaria, K.C. Robbins, Comparison of the esterase andhuman plasminogen activator activities of various activated forms of humanplasminogen and their equimolar streptokinase complexes, J. Biol. Chem. 252(1977) 1141–1147.

[33] F. Xue, Z. Gu, J.A. Feng, LINKER: a web server to generate peptide sequences withextended conformation, Nucleic Acids Res. 32 (2004) 562–565.

[34] P.S. Arora, A.Z. Ansari, T.P. Best, M. Ptashne, P.B. Dervan, Design of artificialtranscriptional activators with rigid poly-L-proline linkers, J. Am. Chem. Soc. 124(2002) 13067–13071.