Báo cáo Y học: The role of the second binding loop of the cysteine protease inhibitor, cystatin A (stefin A), in stabilizing complexes with target proteases is exerted predominantly by Leu73 pdf

The role of the second binding loop of the cysteine protease inhibitor, cystatin A (stefin A), in stabilizing complexes with target proteases is exerted predominantly by Leu73 Alona Pavlova, Sergio Estrada* and Ingemar Bjo¨rk Department of Veterinary Medical Chemistry, Swedish University of Agricultural Sciences, Uppsala Biomedical Centre, Sweden The aim of this work was to elucidate the roles of individual residues within the flexible second binding loop of human cystatin A in the inhibition of cysteine proteases. Four recombinant variants of the inhibitor, each with a single mutation, L73G, P74G, Q76G or N77G, in the most exposed part of this loop were generated by PCR-based site- directed mutagenesis. The binding of these variants to papain, cathepsin L, and cathepsin B was characterized by equilibrium and kinetic methods. Mutation of Leu73 decreased the affinity for papain, cathepsin L and cathep- sin B by  300-fold, >10-fold and  4000-fold, respect- ively. Mutation of Pro74 decreased the affinity for cathepsin B by  10-fold but minimally affected the affinity for the other two enzymes. Mutation of Gln76 and Asn77 did not alter the affinity of cystatin A for any of the proteases studied. The decreased affinities were caused exclusively by increased dissociation rate constants. These results show that the second binding loop of cystatin A plays a major role in stabilizing the complexes with proteases by retarding their dissociation. In contrast with cystatin B, only one amino- acid residue of the loop, Leu73, is of principal importance for this effect, Pro74 assisting to a minor extent only in the case of cathepsin B binding. The contribution of the second binding loop of cystatin A to protease binding varies with the protease, being largest,  45% of the total binding energy, for inhibition of cathepsin B. Keywords: cathepsins; cystatin; cysteine proteases; papain; second binding loop. Cystatins are effective protein inhibitors of cysteine pro- teases of the papain superfamily (reviewed in [1–4]). Found both intracellularly and extracellularly, they are believed to control the activity of normal endogenous proteases, as well as to protect organisms from the harmful activity of exogenous cysteine proteases [1,4–11]. They are generally classified into three families according to their size and the presence of internal disulfide bonds. Cystatins of family 1, also called stefins, are small nonglycosylated proteins  11– 12 kDa in size without disulfide bonds. Family 2 cystatins are somewhat larger,  12–14 kDa, with a structure stabi- lized by two disulfide bonds. Kininogens, representing the third family, are glycosylated proteins of about 50–90 kDa. The single polypeptide chain of a kininogen contains three domains resembling family 2 cystatins. Cystatins competitively inhibit the activity of papain- like cysteine proteases by binding to the active site of the latter and forming a tight, reversible protein–protein complex. A model of the inhibition was initially proposed from computer docking experiments based on the X-ray structures of papain and chicken cystatin, a family 2 member [12]. This model was later substantiated by the X-ray structure of a complex of the family 1 cystatin, human cystatin B (stefin B), with papain [13], the only structure of a cystatin–protease complex determined so far. The N-terminal segment and two hairpin loops of the cystatin together form a hydrophobic wedge-shaped edge that fits well into the active-site cleft of papain. The high degree of complementarity between the interacting surfa- ces allows the complex to form without significant conformational changes of either papain or the inhibitor [12–18]. Both the similar three-dimensional structures of cystatins of families 1 and 2 [12,13,19–21] and the pronounced sequence homology and similar fold of cysteine proteases of the papain family [4,11,22–24] indicate that the general aspects of the interaction model can be extended to complexes between cystatins and other members of this protease family. However, certain distin- guishing features of the structures of some cysteine proteases, such as the occluding loop of cathepsin B [25], cause the mode of inhibition to deviate somewhat for these enzymes. Cystatins thus inhibit cathepsin B by a two-step reaction involving displacement of the occluding loop of the protease in the second step [26,27]. Moreover, it is apparent that the role of an individual binding region Correspondence to I. Bjo ¨ rk, Department of Veterinary Medical Chemistry, Swedish University of Agricultural Sciences, Uppsala Biomedical Centre, Box 575, SE-751 23 Uppsala, Sweden. Fax: + 46 18 550762, Tel.: + 46 18 4714191, E-mail: Ingemar.Bjork@vmk.slu.se Abbreviations: app, subscript denoting an apparent equilibrium or rate constant determined in the presence of an enzyme substrate; E-64, 4-[(2S,3S)-3-carboxyoxiran-2-carbonyl- L -leucylamido]butylguani- dine; His-tag, 10 successive histidine residues fused to an expressed protein; k ass , bimolecular association rate constant; K d , dissociation equilibrium constant; k diss , dissociation rate constant; K i ,inhibition constant; k obs , observed pseudo-first-order rate constant. *Present address: PET-Centre, Uppsala University, University Hospital, SE-751 85 Uppsala, Sweden. (Received 12 July 2002, revised 16 September 2002, accepted 20 September 2002) Eur. J. Biochem. 269, 5649–5658 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03273.x of the inhibitor in protease binding can differ with the target protease [28]. The contributions of the N-terminal region and the first binding loop of family 1 and 2 cystatins, as well as of the second binding loop of family 2 cystatins, to the inhibition of cysteine proteases have been elucidated [29–36]. Recent work has also demonstrated the importance of two amino- acid residues, Leu73 and His75, in the second binding loop of the family 1 inhibitor, cystatin B, for high-affinity binding to a number of cysteine proteases [37]. The sequence of the corresponding hairpin loop in cystatin A (stefin A), another member of family 1, is appreciably different from that in cystatin B; in particular, His75 of cystatin B is substituted by Gly in cystatin A [1]. Moreover, the NMR structure of cystatin A shows that the second loop of this inhibitor is highly flexible, which might be expected to affect the interactions with the protease [20]. It is thus unclear whether the second binding loop of cystatin A fulfils the same function as the second binding loops of cystatin B and family 2 cystatins and also what residues of this loop in cystatin A may participate in the interaction. To elucidate the role of the second binding loop of human cystatin A in the inhibition of cysteine proteases, we have characterized the contribution of four individual amino-acid residues within the most exposed region of this loop (from Leu73 to Asn77) to protease binding (see Fig. 1A). Four recombinant cystatin A variants with Gly replacing each of these amino acids were prepared, and their interaction with papain, cathepsin L, and cathepsin B was characterized by equilibrium and kinetic methods. The results clearly show that the second binding loop of cystatin A is important for the stability of complexes with cysteine proteases. Its quantitative role in protease binding varies with the target enzyme, but is especially important for cathepsin B. Leu73, which is highly conserved in family 1 cystatins, makes the predominant contribution of all residues of the loop to the free energy of formation of the enzyme–inhibitor complex. Pro74 is of minimal importance for the interaction with papain and cathepsin L but participates to some extent in cathepsin B binding. However, the roles of Gln76 and Asn77 in the protease inhibition are negligible. MATERIALS AND METHODS Construction of expression vectors for cystatin A second-loop mutants A previously developed expression vector containing the human cystatin A coding sequence preceded by successive sequences for the leader peptide for the outer membrane protein A of Escherichia coli, a His-tag, and the recognition site for enterokinase was used in this work [38]. This vector has a kcl857 temperature-sensitive repressor gene, allowing induction of expression by increasing the temperature, and an ampicillin-resistance gene [18]. Residues Leu73, Pro74, Gln76, and Asn77 within the second binding loop of cystatin A were substituted with Gly by PCR-based site- directed mutagenesis [39]. Briefly, two mutagenic primers and two standard PCR primers, the latter being comple- mentary to regions of the vector flanking the cysta- tin A-coding sequence, were used for creation of each mutant (Table 1). The desired mutation was introduced in two steps. First, two overlapping DNA fragments, bearing Fig. 1. Model of the three-dimensional structure of the complex between cystatin A and active papain. (A) Overall structure of the complex in ribbon representation, with cystatin A in green and papain in blue. Residues in the second binding loop of cystatin A mutated in this work areinred.PapainresiduesinvolvedininteractionswiththecystatinA second-binding-loop residues are in black. (B) Close-up view of the interactions between residues in the second binding loop of cystatin A and papain residues. The colors of the residues are as in (A). Inter- molecular hydrophobic contacts within a distance of 4 A ˚ are repre- sented as dashed lines. The model is derived from the X-ray structure of the human C3S-cystatin B–S-(carboxymethyl)papain complex (PDB entry 1STF) [13]. 5650 A. Pavlova et al.(Eur. J. Biochem. 269) Ó FEBS 2002 the same mutation, were synthesized in two separate PCRs, in each of which a mutagenic and a standard primer were used and the cystatin A expression vector was the template. In the next step, a larger DNA fragment containing the entire mutant cystatin A-coding sequence was obtained by a third PCR with the standard PCR primers and with a mixture of the products of the previous two PCRs as template. The resulting DNA fragment was cleaved with NcoIandBamHI, and the purified cleavage product containing the mutant cystatin A cDNA was cloned into the original vector between the NcoIandBamHI restriction sites, replacing the corresponding region coding for wild- type cystatin A [38]. The vector was then transformed into E. coli strain MC 1061, made competent with CaCl 2 [40], and transformants were selected by growing the bacteria on agar plates containing ampicillin. Plasmids from a number of colonies of each mutant were purified, and those with the correct mutant cystatin A cDNA were identified by sequencing in an ABI PRISMÒ 310 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA). Expression and purification of cystatin A mutants Recombinant L73G, P74G, Q76G, and N77G cystatin A variants were expressed in E. coli essentially as described previously [18]. The recombinant proteins were purified from periplasmic extracts by immobilized metal affinity chromatography on HisBindÒ Resin (Novagen, Madison, WI, USA), charged with Ni 2+ , or Ni/nitrilotriacetate agarose (Qiagen, Hilden, Germany), as in previous work [38]. The His-tag was cleaved off with enterokinase (EC 3.4.21.9; Biozyme Laboratories, Blaenavon, UK), and the liberated cystatin A mutant was isolated by rechromato- graphy on the same affinity column [38]. Intact His-tagged fusion proteins still contaminating some preparations were removed by absorption on a TALON TM Metal Affinity Resin (Clontech, Palo Alto, CA, USA) by a hybrid batch/ gravity flow column procedure according to a protocol from the manufacturer. Chicken cystatin Forms1and2ofchickencystatinwereisolatedfrom chicken egg white [41]. The two forms have the same sequence and are functionally identical [41], although form 2 is phosphorylated at Ser80 [42] and therefore has a lower isoelectric point. Proteases Papain (EC 3.4.22.2) was purified, stored as inactive S-(methylthio)papain and activated before use as in previ- ous work [41]. The thiol group content of the activated papain, determined by reaction with 5,5¢-dithiobis(2-nitro- benzoic acid) [43], was 0.95–1.00 mol per mol of enzyme. Titrations with chicken cystatin (form 1) [41] gave a cystatin to papain stoichiometry of 0.98 ± 0.02, indicating that the enzyme was fully active in binding cystatins. Cathepsin L (EC 3.4.22.15) from sheep liver was a gift from R. W. Mason, Alfred I. du Pont Institute, Wilmington, DE, USA. Human liver cathepsin B (EC 3.4.22.1) was obtained from Calbiochem (San Diego, CA, USA). Determination of protein concentration Most protein concentrations were calculated from A 280 measurements. Molar absorption coefficients of 55 900 M )1 Æcm )1 for papain and S-(methylthio)papain [41], 8800 M )1 Æcm )1 for all forms of cystatin A [18], and 11 400 M )1 Æcm )1 for chicken cystatin [41] were used. The concentration of active cathepsin L was determined by titration with 4-[(2S,3S)-3-carboxyoxiran-2-carbonyl- L -leu- cylamido]butylguanidine (E-64) [44]. The concentration of cathepsin B was provided by the manufacturer. Binding stoichiometries The stoichiometries of binding of the cystatin A variants to papain were determined at least in duplicate by titrations of 1 l M active papain or S-(methylthio)papain with the variants. The binding to active papain was monitored by following the decrease in activity of the enzyme with a chromogenic substrate [38], whereas the binding to S-(methylthio)papain was monitored by following the change in tryptophan fluorescence accompanying the interaction [41]. The binding stoichiometries were deter- mined by nonlinear least-squares regression analysis of the titration curves [41]. Inhibition constants Apparent inhibition constants, K i , app , for the inhibition of cathepsins L and B by the cystatin A mutants were obtained from the equilibrium rates of hydrolysis of a fluorogenic substrate by the enzyme at different inhibitor concentrations Table 1. Primers for construction of expression vectors for cystatin A second-loop mutants. All sequences are given in the 5¢fi3¢ direction. Codons for Gly, replacing residues to be mutated, are underlined, and base changes introducing the mutations are in bold. Primer Mutation Direction Sequence Standard All Forward GCTCAGGCGACCATGGGCCATCATCATC Reverse CTTGCATGCCCTGCAGGTCG Mutagenic L73G Forward GTATTCAAAAGTGGTCCCGGACAAAATGAG GACTTG Reverse TCCGGGACCACTTTTGAATACTTTCAAGTGCATATATTTATT P74G Forward CAAAAGTCTTGGCGGACAAAATGAGGACTTGGTAC Reverse CATTTTGTCCGCCAAGACTTTTGAATACTT TCAAGTGC Q76G Forward CTTCCCGGAGGAAATGAGGACTTGGTACTTACTG Reverse CCTCATTTCCTCCGGGAAGACTTTTGAATA C N77G Forward CGGACAAGGTGAGGACTTGGTACTTACTGGATAC Reverse CAAGTCCTCACCTTGTCCGGGAAGACTTTTG Ó FEBS 2002 Second protease-binding loop of cystatin A (Eur. J. Biochem. 269) 5651 [28,32]. Product formation was continuously monitored in a conventional fluorimeter (F-4000; Hitachi, Tokyo, Japan) as in previous work [28]. The substrates were carbobenz- oxy- L -phenylalanyl- L -arginine 4-methylcoumaryl-7-amide (Peptide Institute, Osaka, Japan) for cathepsin L and carbobenzoxy- L -arginyl- L -arginine 4-methylcoumaryl- 7-amide (Peptide Institute) for cathepsin B at concentra- tions of 5 and 10 l M , respectively. The fluorescence never exceeded that corresponding to 5% substrate hydrolysis. Inhibitor concentrations were at least 10-fold higher than enzyme concentrations. The inhibition of the enzymes by L73G-cystatin A was analysed at cystatin concentrations ranging from (0.1–0.5) · K i,app to (6–10) · K i,app . Corres- ponding measurements with P74G-cystatin A were per- formed at inhibitor concentrations varying from (0.5–2) · K i,app to (10–14) · K i,app , whereas the range was from (3–4) · K i,app to (10–30) · K i,app for Q76G-cystatin A and N77G-cystatin A. Values of K i,app were derived by nonlinear regression analyses of plots of the ratio between the inhibited and uninhibited rates of substrate hydrolysis against inhibitor concentration [32]. True inhibition con- stants, K i , were obtained after correction for substrate competition [32,45,46]. Association kinetics Association rate constants, k ass, for the inhibition of papain and cathepsins L and B by the cystatin A mutants were determined by continuously monitoring the loss of enzyme activity in the presence of a fluorogenic substrate in either a conventional fluorimeter (see above) or a stopped-flow fluorimeter (SX-17 MV ; Applied Biophysics, Leatherhead, UK) [28,38]. The substrate for papain was 10 l M carbo- benzoxy- L -phenylalanyl- L -arginine 4-methylcoumaryl-7- amide (Peptide Institute), and the substrates for cathepsins L and B and their concentrations were the same as those used to determine K i (see above). The fluorescence was always lower than that given by 5% substrate hydrolysis. The concentrations of the inhibitors were at least 10-fold higher than those of the enzymes and were varied in a 10– 20-fold range. The highest inhibitor concentrations in reactions with papain and cathepsin L were 10–20 n M , whereas reactions with cathepsin B were analyzed at inhibitor concentrations up to 30 l M for L73G-cystatin A andupto0.3–0.5l M for the other mutants. Apparent pseudo-first-order rate constants, k obs,app , were obtained by nonlinear least-squares regression analysis of the progress curves [28]. Apparent association rate constants, k ass,app , were calculated from the slopes of plots of k obs,app vs. inhibitor concentration and were corrected for substrate competition to give the true association rate constants, k ass [28,45–47]. Dissociation kinetics Dissociation rate constants, k diss , for the complexes of the cystatin A mutants with papain were determined by dis- placement experiments, essentially as detailed previously [14,16]. Papain dissociating from the complexes was trapped by a high excess of chicken cystatin (form 2), which binds faster and more tightly to papain than cystatin A or the cystatin A mutants do [14,18] (see also Results) and thereby prevents reassociation of the cystatin A variants with the enzyme. The concentration of the cystatin A mutant– papain complexes was 2.5–5.0 l M , and the molar ratio of the displacing chicken cystatin to the complexes varied between 10-fold and 50-fold. The progress of the reaction was monitored for 100–150 h by following the appearance of the newly formed complex between papain and chicken cystatin, analyzed by ion-exchange chromatography on a MonoQ TM column (Amersham Biosciences, Uppsala, Sweden). Form 2 of chicken cystatin was used because its lower isoelectric point allows the complex with papain to be well separated and thus easily quantified in this analysis. k diss was calculated as described previously [14]. k diss for the complex between L73G-cystatin A and cathepsin L was determined by trapping the enzyme dissociated from the complex by a high concentration of the substrate, carbobenzoxy- L -phenylalanyl- L -arginine 4-methylcoumaryl-7-amide, which binds tightly to cathep- sin L with a K m of 1.8 l M [45]. In most experiments, the complex was formed by incubating 0.04 n M cathepsin L with 0.4 n M L73G-cystatin A for 90 min, which resulted in an essentially complete reaction, with  80% of the enzyme being saturated with the inhibitor. The substrate was then added to a final concentration of 100 l M with minimal dilution of the complex. Alternatively, the complex was formed by incubation of 1 n M cathepsin L with 10 n M L73G-cystatin A for 15 min, resulting in  99% of the enzyme being bound in the complex, and this mixture was then diluted 1000-fold into 100 l M substrate. In both cases, the dissociation of the complex was monitored in a conventional fluorimeter by continuously recording the fluorescence increase due to cleavage of the substrate by the liberated cathepsin L. The fluorescence never exceeded that corresponding to 5% substrate hydrolysis. k diss was deter- mined by nonlinear least-squares regression analysis of the progress curves [15]. Fluorescence emission spectroscopy Fluorescence emission spectra of free papain and wild-type or L73G-cystatin A, as well as of complexes of papain with either of the two cystatin A variants, were recorded in an SLM 4800S spectrofluorimeter (SLM-Aminco, Urbana, IL, USA) with an excitation wavelength of 280 nm, as described previously [16,41]. Papain and cystatin concen- trations were 1.0 and 1.2 l M , respectively, giving > 99% saturation of enzyme with inhibitor in analyses of the complexes. All spectra were corrected for inner-filter effects and for the wavelength dependence of the instrumental response [41] and were normalized to a fluorescence intensity of 1.0 for free papain at the wavelength of the emission maximum. Difference spectra between the com- plexes and the free proteins were calculated as in [41]. Protein modeling The structure of human cystatin A in complex with active papain was modeled on to the X-ray structure of the complex between human C3S-cystatin B and S-(carboxy- methyl)papain (PDB entry 1STF) [13] with the program SWISS - PDB Viewer (http://www.expasy.ch/spdbv/). The most favorable rotamers of the side chains of the 46 residues of cystatin A which differ from those of cystatin B [1] were initially selected by the program [48], and the 5652 A. Pavlova et al.(Eur. J. Biochem. 269) Ó FEBS 2002 S-carboxymethyl group of the papain moiety of the complex was removed in the same manner. The model was then corrected by the program facility ÔQuick and Dirty FixingÕ of all side chains in the complex, followed by ÔExhaustive Search FixingÕ of the side chains within the Leu73–Asn77 segment in the second binding loop of cystatin A. After each of these steps, the conformation of the second binding loop in the complex was refined by energy minimization of the Leu73–Asn77 segment and all neighboring residues within 6 A ˚ . The possibility of other residues replacing Gly75 in the final model was evaluated by ÔQuick and Dirty FixingÕ of all side chains in the complex after each replacement. Miscellaneous procedures For N-terminal sequencing and determination of molecular masses, the mutants were first desalted into 0.1% (v/v) trifluoroacetic acidby gel chromatographyon Fast-Desalting PC 3.2/10 columns (Amersham Biosciences). N-Terminal sequences were analyzed by Edman degradation in an Applied Biosystems 477A Protein Sequencer. Molecular masses were measured by MALDI MS in a Kratos Kompact MALDI 4 instrument (Kratos, Manchester, UK) as in [18]. SDS/PAGE under reducing and nonreducing condi- tions was performed with the Tricine buffer system [49]. Experimental conditions All equilibrium and kinetic experiments were performed at 25.0 ± 0.2 °C. The proteases were first activated by 1 m M dithiothreitol in the reaction buffer for 10 min at 25 °C. The inhibition of papain was studied in 50 m M Tris/HCl, pH 7.4, containing 100 m M NaCl, 0.1 m M EDTA and, except in the displacement experiments, 1 m M dithiothreitol and 0.01% (w/v) BrijÒ 35. The interaction with cathepsin L wasanalyzedin100 m M sodium acetate, pH 5.5, containing 100 m M NaCl, 1 m M EDTA, 1 m M dithiothreitol, and 0.01% (w/v) BrijÒ 35, whereas the buffer for cathepsin B was 50 m M Mes/NaOH, pH 6.0, containing 100 m M NaCl, 0.1 m M EDTA, 1 m M dithiothreitol, and 0.1% (w/v) poly(ethylene glycol) 6000. RESULTS Preparation, homogeneity and activity of cystatin A mutants Four variants of cystatin A, each with a single amino-acid residue, Leu73, Pro74, Gln76 or Asn77, substituted by Gly were produced by recombinant DNA techniques. All these mutations are in the most exposed part of the second protease-binding loop of the inhibitor (Fig. 1A). Residue 75 was not substituted, as it is Gly in the wild-type sequence. The expression vectors were constructed by PCR-based site- directed mutagenesis and contained the expected mutant sequences in the case of the L73G, P74G and N77G mutants. However, all vectors for the Q76G mutant purified from 18 individual clones had, in addition to the desired mutation, a T fi C substitution in the codon for Thr83. This substitution was in the region specified by the forward mutagenic primer for this mutant and was probably due to an erroneously synthesized primer. As this additional mutation is silent, one of the isolated vectors was neverthe- less used for expression of Q76G-cystatin A. The mutants were expressed with a removable His-tag and with a signal peptide directing the proteins to the periplasmic space of E. coli, facilitating purification. All purified mutants were > 99.5% homogeneous on SDS/PAGE. N-Terminal sequencing of the first five residues confirmed that the His-tag was cleaved off properly by enterokinase for all mutants. The molecular masses, determined by MS, corresponded within 4.5 Da to those calculated from the expected amino-acid sequences, con- firming the correct length of the mutants, as well as the presence of the desired mutations. All mutants bound active papain and S-(methylthio)papain with stoichiometries between 0.95 and 1.0, i.e. they were essentially fully active in inhibition of cysteine proteases. Binding affinity All four cystatin A mutants bound so tightly to papain that the affinity of the binding could not be determined by equilibrium methods, because of the instability of the enzyme at the low concentrations and the long reaction times that would have been necessary. Therefore, K d for the interaction with papain was calculated as k diss /k ass from independently measured rate constants (see below and Table 2), as was K d for wild-type cystatin A binding to this enzyme in previous work [18]. Only the L73G mutation caused a pronounced,  300-fold, decrease in the affinity for papain, compared with that of the wild-type inhibitor (Table 2). In contrast, the P74G, Q76G, and N77G muta- tions resulted in minimal, less than twofold, changes in affinity. The high affinity of most mutants for cathepsin L also precluded an accurate determination of K i from equilibrium measurements. Such experiments gave only upper limits of K i for the interaction of P74G, Q76G, and N77G cystatin A with this protease (Table 2), similar to previous analyses of K i for the wild-type inhibitor [35]. Moreover, as k diss for these tight interactions could not be determined (see below), K d could not be calculated from the rate constants. No meaningful comparisons of the affinities of the three mutants for cathepsin L with that of wild-type cystatin A were therefore possible. However, a reliable K i for the inhibition of cathepsin L by L73G-cystatin A was obtained by equilibrium measurements and was > 10-fold higher than that for wild-type cystatin A. The measured K i for this mutant agreed well with K d calculated from k ass and k diss (see below and Table 2). K i for the inhibition of cathepsin B by all cystatin A forms was sufficiently high to be well determined by equilibrium analyses. The L73G mutation caused a sub- stantial,  4000-fold, increase in K i which was confirmed by calculations of K d from k ass and k diss (Table 2). A smaller,  10-fold, increase in K i for cathepsin B was also observed for the P74G mutant, whereas the affinities of both Q76G and N77G cystatin A for the enzyme differed minimally, about twofold, from that of wild-type cystatin A (Table 2). Association rate constants The kinetics of association of the cystatin A mutants with papain, cathepsin L and cathepsin B were analyzed Ó FEBS 2002 Second protease-binding loop of cystatin A (Eur. J. Biochem. 269) 5653 by continuously monitoring the decrease in enzyme acti- vity against a fluorogenic substrate. Most reactions were studied by conventional fluorimetry, whereas the rapid association of L73G-cystatin A with cathepsin B required the use of stopped-flow measurements. All progress curves were well fitted to a single-exponential function. Plots of the dependence of k obs,app , derived from these fits, on inhibitor concentration were linear in the concen- tration range covered for all mutants. Values of k ass were determined from the slopes of these plots. All four amino-acid substitutions in the second binding loop of cystatin A had a marginal effect on k ass for the binding to papain, cathepsin L or cathepsin B (Table 2). Dissociation rate constants The low dissociation rate constants of the complexes between the cystatin A mutants and papain were measured by displacement of the mutants from the complexes with an excess of a tighter-binding inhibitor, chicken cystatin, in experiments monitored by ion-exchange chromatography. Only the L73G mutation altered k diss to any appreciable extent, increasing it by  170-fold over that for wild-type cystatin A. k diss for all other mutants was essentially unaffected, with at most a twofold increase being observed for P74G-cystatin A. k diss of the complex between L73G-cystatin A and cathepsin L was measured by displacement experiments, in which the enzyme dissociating from the complex was cap- tured by an excess of a tight-binding fluorogenic substrate. The values of k diss obtained by two modifications of this procedure agreed well with each other and with that calculated from K i and k ass (Table 2). The L73G mutation resulted in a greater than sevenfold increase in k diss , compared with the value for wild-type cystatin A. This method could not be used to determine k diss for the complexes of the P74G, Q76G, and N77G mutants with cathepsin L, because of the high stabilities of these complexes and, consequently, very long dissociation times. Moreover, the limited amounts of cathepsin L available precluded Table 2. Equilibrium and rate constants at 25 °C for the binding of cystatin A variants with substitutions in the second binding loop to papain, cathepsin L and cathepsin B. Methods and experimental conditions are described in Materials and Methods. Values determined in this work are given as means ± SEM with the number of experiments in parentheses. Values for wild-type cystatin A, reported previously and shown for comparison, as well as calculated values are given without errors. Numbers in square brackets indicate the ratio of the corresponding constant to that for wild-type cystatin A. Enzyme Cystatin A form K d ( M ) k ass ( M )1 Æs )1 ) k diss (s )1 ) Papain Wild-type 1.8 · 10 )13 a 3.1 · 10 6a 5.5 · 10 )7a [1] [1] [1] L73G 5.8 · 10 )11 b (1.58 ± 0.02) · 10 6 (9) (9.1 ± 0.9) · 10 )5 (3) [320] [0.5] [170] P74G 2.8 · 10 )13 b (3.64 ± 0.06) · 10 6 (9) (10.2 ± 0.7) · 10 )7 (3) [1.6] [1.2] [1.9] Q76G 1.8 · 10 )13 b (3.07 ± 0.02) · 10 6 (8) (5.6 ± 0.5) · 10 )7 (3) [1] [1] [1] N77G 0.95 · 10 )13 b (3.58 ± 0.07) · 10 6 (9) (3.4 ± 0.3) · 10 )7 (3) [0.5] [1.2] [0.6] Cathepsin L Wild-type £ 1 · 10 )11 a 5.2 · 10 6a £ 5 · 10 )5a [1] [1] [1] L73G (1.09 ± 0.08) · 10 )10 (10) (2.98 ± 0.04) · 10 6 (10) (3.4 ± 0.4) · 10 )4 (3) [‡ 11] [0.6] [‡ 7] 1.1 · 10 )10 b 3.2 · 10 )4c P74G £ 2.4 · 10 )11 (7) (4.6 ± 0.2) · 10 6 (10) [0.9] £ 1.1 · 10 )4c Q76G £ 1.1 · 10 )11 (8) (6.3 ± 0.2) · 10 6 (9) [1.2] £ 6.9 · 10 )5c N77G £ 1.1 · 10 )11 (8) (6.3 ± 0.2) · 10 6 (14) [1.2] £ 6.9 · 10 )5c Cathepsin B Wild-type 9.1 · 10 )10 a 3.9 · 10 4a 3.5 · 10 )5a [1] [1] [1] L73G (3.6 ± 0.2) · 10 )6 (11) (8.5 ± 0.3) · 10 4 (8) 0.28 ± 0.05 (8) [4000] [2.2] [8000] 3.3 · 10 )6b 0.31 c P74G (9.7 ± 0.6) · 10 )9 (10) (5.5 ± 0.2) · 10 4 (11) 5.3 · 10 )4c [11] [1.4] [15] Q76G (1.4 ± 0.1) · 10 )9 (8) (3.95 ± 0.07) · 10 4 (8) 5.5 · 10 )5c [1.5] [1] [1.6] N77G (2.4 ± 0.2) · 10 )9 (9) (2.26 ± 0.08) · 10 4 (11) 5.4 · 10 )5c [2.6] [0.6] [1.5] a From previous work [18,35]. b Calculated from k ass and k diss . c Calculated from K i and k ass . 5654 A. Pavlova et al.(Eur. J. Biochem. 269) Ó FEBS 2002 determination of k diss by the method used for papain, in which the displacement was monitored by chromatography. Therefore, only upper limits of k diss for the binding of these mutants to cathepsin L could be estimated, as for wild-type cystatin A in previous work [35] (Table 2). Values of k diss for the complexes of the four cystatin A mutants with cathepsin B were calculated from K i and k ass determined in separate experiments (Table 2). In addition, k diss for the L73G-cystatin A–cathepsin B complex was obtained from the analyses of the association kinetics (see above) as the intercept on the ordinate of the plot of k obs,app vs. inhibitor concentration and was in a good agreement with the calculated k diss (Table 2). The L73G mutation markedly affected the rate of dissociation of the complex with cathepsin B, increasing k diss by  8000-fold. The P74G mutation resulted in a smaller,  15-fold, increase in k diss , whereas the other mutations altered k diss minimally. Fluorescence emission difference spectra The fluorescence difference spectra between complexes of wild-type or L73G-cystatin A with papain and the free proteins had minima at different wavelengths,  360 and  368 nm, respectively (Fig. 2). Moreover, the spectrum for the L73G mutant had an appreciably lower amplitude than that for wild-type cystatin A, reflecting a smaller fluores- cence change on interaction of the mutant than of the wild- type inhibitor with papain. These fluorescence changes must reflect different changes in the environment of one or more Trp side chains in papain on formation of the two enzyme– inhibitor complexes, as cystatin A does not contain Trp [1]. DISCUSSION The X-ray structure of the complex between human C3S- cystatin B and S-(carboxymethyl)papain reveals a number of predominantly hydrophobic but also solvent-mediated interactions between the second binding loop of the inhibitor and papain [13]. In particular, Leu73 and His75 in this loop are seen to make four and seven intermolecular contacts with the enzyme, respectively, that are < 4 A ˚ in length. In agreement with this structural evidence, site- directed mutagenesis has shown that the two residues contribute substantial free energy to the interaction of cystatin B with cysteine proteases [37]. The sequence of the second binding loop of the related family 1 cystatin, cystatin A, differs appreciably from that of cystatin B. Most notably, cystatin A lacks the essential His75 and instead has a Gly in this position [1]. This substitution would be expected to lead to loss of a number of interactions with the enzyme and therefore to considerably decrease the contribution of the second binding loop of cystatin A to the inhibition of target proteases. However, the flexibility of the second binding loop of cystatin A demonstrated by NMR [20] may allow other amino acids of the loop to make additional favorable contacts with the protease, thereby compensating for the absence of this residue. Alternatively, this flexibility may instead destabilize the interactions of the loop with the target protease, resulting in a lower binding energy. To clarify the functional role of the second binding loop of cystatin A, we have studied the contribution by the residues within the most exposed segment of this loop to the inhibition of cysteine proteases. The L73G, P74G, Q76G, and N77G mutants of cystatin A were constructed by site- directed mutagenesis, and their inhibition of papain, cathepsin L and cathepsin B was characterized. Our results show that the second binding loop of cystatin A is essential for the formation of tight complexes between the inhibitor and the cysteine proteases studied. However, in contrast with cystatin B, this role is exerted predominantly by only one residue, Leu73, which is highly conserved in family I cystatins. The major role of Leu73 in the interactions is in stabilizing the complexes once they are formed. This conclusion is indicated by the L73G mutation appreciably decreasing the affinity of cystatin A for the proteases by increasing the rate constants for dissociation of the complexes but negligibly affecting the association rate constants. This contribution of Leu73 to the inhibitory ability of cystatin A varies for different target proteases, being most pronounced for the inhibition of cathepsin B. The Leu73 side chain thus contributes about )15 and )21 kJÆmol )1 to the unitary free energy change [50,51] accompanying the formation of the complex of cystatin A with papain and cathepsin B, respectively. These changes correspond to  18 and  34%, respectively, of the total unitary free energy of binding of cystatin A to the two enzymes [18]. The contribution of Leu73 of cystatin A to binding of papain, which has an open active-site cleft, is comparable to that of Leu73 in the second binding loop of cystatin B and to that of the essential Trp106 residue in this loop of the family 2 cystatin, cystatin C [37,52]. However, the contribution of Leu73 of cystatin A to binding of cathepsin B, in which the occluding loop partially blocks the active site, is substantially higher than that of the corres- ponding residue of cystatin C [52]. The results further show that one additional residue in the second binding loop of cystatin A, Pro74, aids in stabilizing the complex of the inhibitor with cathepsin B by decreasing the dissociation rate constant. However, this Fig. 2. Fluorescence emission difference spectra between complexes of human wild-type cystatin A or the L73G cystatin A variant with papain andthefreeproteins.Solid line, Wild-type cystatin A; dotted line, L73G-cystatin A. Fluorescence emission spectra were measured as describedinMaterialsandmethodswithpapainandcystatincon- centrations of 1.0 and 1.2 l M , respectively. The difference spectra were calculated from separately measured and corrected emission spectra that were normalized to a fluorescence intensity of 1.0 for 1 l M papain at the wavelength of the emission maximum [41]. Ó FEBS 2002 Second protease-binding loop of cystatin A (Eur. J. Biochem. 269) 5655 residue negligibly participates in the inhibition of papain and most likely also of cathepsin L. The side chains of Leu73 and Pro74 jointly contribute  45% of the total unitary free energy of binding of cystatin A to cathepsin B, demonstrating a major role of the second binding loop of cystatin A in the inhibition of this enzyme. Pro74 may be directly involved in the interaction with cathepsin B by providing hydrophobic interactions with the protease. Alternatively, the role of Pro74 might be to maintain an appropriate orientation of Leu73 for its specific interaction with cathepsin B. In contrast with Leu73 and Pro74, the two other residues of the second binding loop of cystatin A studied, Gln76 and Asn77, are of minimal importance for the affinity of the inhibitor for the cysteine proteases and therefore presumably do not interact directly with the enzymes. The remaining residue of the loop, Gly75, may conceivably provide backbone interactions with a target protease, but such a contribution cannot be investigated by the approach taken in this work. The conclusions drawn above from the results of this work are in general agreement with modeling of the cystatin A–papain complex. Although no X-ray structure of cystatin A is available, the NMR structure of the inhibitor is similar to the X-ray structure of human C3S-cystatin B in complex with S-(carboxymethyl)papain [13,20,53]. More- over, human cystatins A and B are homologous, having identical amino acids in 52 out of 98 positions [1]. Cystatin B in the complex with papain can therefore be used as an appropriate template for modeling of the corresponding complex between cystatin A and this prote- ase with reasonable accuracy [48,54]. The model generated for the complex indicates that only two residues within the second binding loop of cystatin A, Leu73 and Pro74, are involved in interactions with papain (Fig. 1B). Leu73 makes six hydrophobic interactions of 3.4–4.0 A ˚ with Trp177 of papain in the model, in agreement with the demonstration that Leu73 is essential for strong inhibition of cysteine proteases by cystatin A. The involvement of Trp177 of papain in the interaction with Leu73 is supported by the changes caused by the L73G mutation of the fluorescence difference spectrum characterizing the cystatin A–papain interaction. These changes indicate that one or more tryptophans of papain, probably primarily Trp177 on the surface of the active-site cleft, are exposed to a less hydrophobic environment in the complex with L73G- cystatin A [55]. In the model, the side chain of Pro74 of cystatin A also makes two hydrophobic contacts of  4A ˚ with Gln142 and Leu143 of papain (Fig. 1B). This obser- vation is in apparent contrast with the demonstration that Pro74 is unimportant for papain binding and participates only in the inhibition of cathepsin B. This discrepancy with the experimental data thus indicates that the model is somewhat uncertain with regard to the putative interactions involving Pro74. However, in agreement with the experi- mental results, the model accurately predicts that neither Gln76 nor Asn77 of the second binding loop of cystatin A interact with papain in the complex. Moreover, although the role of Gly75 was not investigated experimentally, the model suggests that Gly is not the only residue in this position compatible with high-affinity interaction with papain. The phi and psi angles of Gly75 deduced from the model are thus within the Ramachandran plot region sterically allowed for other types of residues. Most residues other than Gly could also be modeled into this position without observably interfering sterically with the interac- tion. In conclusion, this study shows the importance of the second binding loop of cystatin A for the binding of cysteine proteases, in particular cathepsin B. The role of this loop is comparable to that of the corresponding loops of cystatin B and family 2 cystatins, to stabilize the cystatin– protease complex by decreasing the dissociation rate. How- ever, in contrast with the latter inhibitors, this role is exerted almost exclusively by one residue of the loop, Leu73, although Pro74 is also of some importance for cathepsin B binding. ACKNOWLEDGEMENTS We are grateful to Dr A ˚ ke Engstro ¨ m (Department of Medical Biochemistry and Microbiology, Uppsala University) for molecular mass determinations and amino-acid sequencing. This project was supported by the Swedish Medical Research Council (Project No. 4212). REFERENCES 1. Barrett, A.J., Rawlings, N.D., Davies, M.E., Machleidt, W., Salvesen, G. & Turk, V. (1986) Cysteine proteinase inhibitors of the cystatin superfamily. In Proteinase Inhibitors (Barrett, A.J. & Salvesen, G., eds), pp. 515–569. Elsevier, Amsterdam. 2. Turk, V. & Bode, W. (1991) The cystatins: protein inhibitors of cysteine proteinases. FEBS Lett. 285, 213–219. 3. Abrahamson, M. (1994) Cystatins. Methods Enzymol. 244,685– 700. 4. Turk,B.,Turk,V.&Turk,D.(1997)Structuralandfunctional aspects of papain-like cysteine proteinases and their protein inhibitors. Biol. Chem. 378, 141–150. 5. Henskens, Y.M.C., Veerman, E.C.I. & Nieuw Amerongen, A.V. (1996) Cystatins in health and disease. Biol. Chem. Hoppe-Seyler 377, 71–86. 6. Blankenvoorde, M.F., Van’t Hof, W., Walgreen-Weterings, E., van Steenbergen, T.J., Brand, H.S., Veerman, E.C. & Nieuw Amerongen, A.V. (1998) Cystatin and cystatin-derived peptides have antibacterial activity against the pathogen Porphyromonas gingivalis. Biol. Chem. 379, 1371–1375. 7. Collins, A.R. & Grubb, A. (1998) Cystatin D, a natural salivary cysteine protease inhibitor, inhibits coronavirus replication at its physiologic concentration. Oral Microbiol. Immunol. 13, 59–61. 8. Kos, J. & Lah, T.T. (1998) Cysteine proteinases and their endogenous inhibitors: target proteins for prognosis, diagnosis and therapy in cancer. Oncol. Rep. 5, 1349–1361. 9. Das, L., Datta, N., Bandyopadhyay, S. & Das, P.K. (2001) Suc- cessful therapy of lethal murine visceral leishmaniasis with cystatin involves up-regulation of nitric oxide and a favorable T cell response. J. Immunol. 166, 4020–4028. 10. Ruzindana-Umunyana, A. & Weber, J.M. (2001) Interactions of human lacrimal and salivary cystatins with adenovirus endo- peptidase. Antiviral Res. 51, 203–214. 11. Turk,V.,Turk,B.&Turk,D.(2001)Lysosomalcysteinepro- teases: facts and opportunities. EMBO J. 20, 4629–4633. 12. Bode, W., Engh, R., Musil, D., Thiele, U., Huber, R., Karshikov, A.,Brzin,J.,Kos,J.&Turk,V.(1988)The2.0A ˚ X-ray crystal structure of chicken egg white cystatin and its possible mode of interaction with cysteine proteinases. EMBO J. 7, 2593–2599. 13.Stubbs,M.T.,Laber,B.,Bode,W.,Huber,R.,Jerala,R., Lenarcic, B. & Turk, V. (1990) The refined 2.4 A ˚ X-ray crystal structure of recombinant human stefin B in complex with the cysteine proteinase papain: a novel type of proteinase inhibitor interaction. EMBO J. 9, 1939–1947. 5656 A. Pavlova et al.(Eur. J. Biochem. 269) Ó FEBS 2002 14. Bjo ¨ rk, I., Alriksson, E. & Ylinenja ¨ rvi, K. (1989) Kinetics of binding of chicken cystatin to papain. Biochemistry 28, 1568–1573. 15. Bjo ¨ rk, I. & Ylinenja ¨ rvi, K. (1990) Interaction between chicken cystatin and the cysteine proteinases actinidin, chymopapain A, and ficin. Biochemistry 29, 1770–1776. 16. Lindahl, P., Abrahamson, M. & Bjo ¨ rk, I. (1992) Interaction of recombinant human cystatin C with the cysteine proteinases papain and actinidin. Biochem. J. 281, 49–55. 17. Turk, B., Colic, A., Stoka, V. & Turk, V. (1994) Kinetics of inhibition of bovine cathepsin S by bovine stefin B. FEBS Lett. 339, 155–159. 18. Pol,E.,Olsson,S.L.,Estrada,S.,Prasthofer,T.W.&Bjo ¨ rk, I. (1995) Characterization by spectroscopic, kinetic and equilibrium methods of the interaction between recombinant human cystatin A (stefin A) and cysteine proteinases. Biochem. J. 311, 275–282. 19. Dieckmann,T.,Mitschang,L.,Hofmann,M.,Kos,J.,Turk,V., Auerswald, E.A., Jaenicke, R. & Oschkinat, H. (1993) The structures of native phosphorylated chicken cystatin and of a recombinant unphosphorylated variant in solution. J. Mol. Biol. 234, 1048–1059. 20. Martin, J.R., Craven, C.J., Jerala, R., Kroon-Zitko, L., Zerovnik, E., Turk, V. & Waltho, J.P. (1995) The three-dimensional solution structure of human stefin A. J. Mol. Biol. 246, 331–343. 21. Ekiel, I., Abrahamson, M., Fulton, D.B., Lindahl, P., Storer, A.C., Levadoux, W., Lafrance, M., Labelle, S., Pomerleau, Y., Groleau, D., LeSauteur, L. & Gehring, K. (1997) NMR structural studies of human cystatin C dimers and monomers. J. Mol. Biol. 271, 266–277. 22. Rawlings, N.D. & Barrett, A.J. (1994) Families of cysteine pep- tidases. Methods Enzymol. 244, 461–486. 23. Turk, D., Guncar, G., Podobnik, M. & Turk, B. (1998) Revised definition of substrate binding sites of papain-like cysteine pro- teases. Biol. Chem. 379, 137–147. 24. Turk, B., Turk, D. & Turk, V. (2000) Lysosomal cysteine proteases: more than scavengers. Biochim. Biophys. Acta 1477, 98–111. 25. Musil, D., Zucic, D., Turk, D., Engh, R.A., Mayr, I., Huber, R., Popovic,T.,Turk,V.,Towatari,T.,Katunuma,N.&Bode,W. (1991) The refined 2.15 A ˚ X-ray crystal structure of human liver cathepsin B: the structural basis for its specificity. EMBO J. 10, 2321–2330. 26. Nycander,M.,Estrada,S.,Mort,J.S.,Abrahamson,M.&Bjo ¨ rk, I. (1998) Two-step mechanism of inhibition of cathepsin B by cystatin C due to displacement of the proteinase occluding loop. FEBS Lett. 422, 61–64. 27. Pavlova,A.,Krupa,J.C.,Mort,J.S.,Abrahamson,M.&Bjo ¨ rk, I. (2000) Cystatin inhibition of cathepsin B requires dislocation of the proteinase occluding loop. Demonstration by release of loop anchoring through mutation of His110. FEBS Lett. 487, 156–160. 28. Bjo ¨ rk,I.,Pol,E.,Raub-Segall,E.,Abrahamson,M.,Rowan, A.D. & Mort, J.S. (1994) Differential changes in the association and dissociation rate constants for binding of cystatins to target proteinases occurring on N-terminal truncation of the inhibitors indicate that the interaction mechanism varies with different enzymes. Biochem. J. 299, 219–225. 29. Machleidt, W., Thiele, U., Laber, B., Assfalg-Machleidt, I., Esterl, A., Wiegand, G., Kos, J., Turk, V. & Bode, W. (1989) Mechanism of inhibition of papain by chicken egg white cystatin. Inhibition constants of N-terminally truncated forms and cyanogen bromide fragments of the inhibitor. FEBS Lett. 243, 234–238. 30. Machleidt, W., Thiele, U., Assfalg-Machleidt, I., Forger, D. & Auerswald, E.A. (1991) Molecular mechanism of inhibition of cysteine proteinases by their protein inhibitors: kinetic studies with natural and recombinant variants of cystatins and stefins. Biomed. Biochim. Acta 50, 613–620. 31. Abrahamson, M., Mason, R.W., Hansson, H., Buttle, D.J., Grubb, A. & Ohlsson, K. (1991) Human cystatin C. Role of the N-terminal segment in the inhibition of human cysteine proteinases and in its inactivation by leucocyte elastase. Biochem. J. 273, 621–626. 32. Lindahl, P., Nycander, M., Ylinenja ¨ rvi,K.,Pol,E.&Bjo ¨ rk, I. (1992) Characterization by rapid-kinetic and equilibrium methods of the interaction between N-terminally truncated forms of chicken cystatin and the cysteine proteinases papain and actinidin. Biochem. J. 286, 165–171. 33. Hall, A., Ha ˚ kansson, K., Mason, R.W., Grubb, A. & Abra- hamson, M. (1995) Structural basis for the biological specificity of cystatin C. Identification of leucine 9 in the N-terminal binding region as a selectivity-conferring residue in the inhibition of mammalian cysteine peptidases. J. Biol. Chem. 270, 5115–5121. 34. Auerswald, E.A., Na ¨ gler, D.K., Assfalg-Machleidt, I., Stubbs, M.T., Machleidt, W. & Fritz, H. (1995) Hairpin loop mutations of chicken cystatin have different effects on the inhibition of cathe- psin B, cathepsin L and papain. FEBS Lett. 361, 179–184. 35. Estrada, S., Pavlova, A. & Bjo ¨ rk, I. (1999) The contribution of N-terminal region residues of cystatin A (stefin A) to the affinity and kinetics of inhibition of papain, cathepsin B, and cathepsin L. Biochemistry 38, 7339–7345. 36. Pol, E. & Bjo ¨ rk, I. (2001) Role of the single cysteine residue, Cys 3, of human and bovine cystatin B (stefin B) in the inhibition of cysteine proteinases. Protein Sci. 10, 1729–1738. 37. Pol, E. & Bjo ¨ rk, I. (1999) Importance of the second binding loop and the C-terminal end of cystatin B (stefin B) for inhibition of cysteine proteinases. Biochemistry 38, 10519–10526. 38. Estrada, S., Nycander, M., Hill, N.J., Craven, C.J., Waltho, J.P. & Bjo ¨ rk, I. (1998) The role of Gly-4 of human cystatin A (stefin A) in the binding of target proteinases. Characterization by kinetic and equilibrium methods of the interactions of cystatin A Gly-4 mutants with papain, cathepsin B, and cathepsin L. Biochemistry 37, 7551–7560. 39. Higuchi, R., Krummel, B. & Saiki, R.K. (1988) A general method of in vitro preparation and specific mutagenesis of DNA frag- ments: study of protein and DNA interactions. Nucleic Acids Res. 16, 7351–7367. 40. Cohen, S.N., Chang, A.C.Y. & Hsu, L. (1972) Nonchromosomal antibiotic resistance in bacteria: genetic transformation of Escherichia coli by R factor DNA. Proc. Natl. Acad. Sci. USA 69, 2110–2114. 41. Lindahl, P., Alriksson, E., Jo ¨ rnvall, H. & Bjo ¨ rk, I. (1988) Inter- action of the cysteine proteinase inhibitor chicken cystatin with papain. Biochemistry 27, 5074–5082. 42. Laber, B., Krieglstein, K., Henschen, A., Kos, J., Turk, V., Huber, R. & Bode, W. (1989) The cysteine proteinase inhibitor chicken cystatin is a phosphoprotein. FEBS Lett. 248, 162–168. 43. Ellman, G.L. (1959) Tissue sulfhydryl groups. Arch. Biochem. Biophys. 82, 70–77. 44. Barrett, A.J. & Kirschke, H. (1981) Cathepsin B, cathepsin H, and cathepsin L. Methods Enzymol. 80, 535–561. 45. Mason, R.W. (1986) Species variants of cathepsin L and their immunological identification. Biochem. J. 240, 285–288. 46. Dalet-Fumeron, V., Guinec, N. & Pagano, M. (1991) High- performance liquid chromatographic method for the simulta- neous purification of cathepsins B, H and L from human liver. J. Chromatogr. 568, 55–68. 47. Hall, A., Abrahamson, M., Grubb, A., Trojnar, J., Kania, P., Kasprzykowska, R. & Kasprzykowski, F. (1992) Cystatin C based peptidyl diazomethanes as cysteine proteinase inhibitors: influence of the peptidyl chain length. J. Enzyme Inhibition 6, 113–123. 48. Guex, N. & Peitsch, M.C. (1997) Swiss-Model and Swiss-Pdb Viewer: an environment for comparative protein modeling. Electrophoresis 18, 2714–2723. 49. Scha ¨ gger, H. & von Jagow, G. (1987) Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal. Biochem. 166, 368–379. Ó FEBS 2002 Second protease-binding loop of cystatin A (Eur. J. Biochem. 269) 5657 50. Gurney, R.W. (1953) Ionic Processes in Solution, pp. 90–105. McGraw-Hill, New York. 51. Karush, F. (1962) Immunologic specificity and molecular struc- ture. Adv. Immunol. 2, 1–40. 52. Bjo ¨ rk, I., Brieditis, I., Raub-Segall, E., Pol, E., Ha ˚ kansson, K. & Abrahamson, M. (1996) The importance of the second hairpin loop of cystatin C for proteinase binding. Characterization of the interaction of Trp-106 variants of the inhibitor with cysteine proteinases. Biochemistry 35, 10720–10726. 53. Craven, C.J., Baxter, N.J., Murray, E.H., Hill, N.J., Martin, J.R., Ylinenja ¨ rvi, K., Bjo ¨ rk, I., Waltho, J.P. & Murray, I.A. (2000) Wild-type and Met-65 fi Leu variants of human cystatin A are functionally and structurally identical. Biochemistry 39, 15783– 15790. 54. Guex, N., Diemand, A. & Peitsch, M.C. (1999) Protein modelling for all. Trends Biochem. Sci. 24, 364–367. 55. Lakowicz, J.R. (1983) Principles of Fluorescence Spectroscopy,pp. 187–214. Plenum Press, New York. 5658 A. Pavlova et al.(Eur. J. Biochem. 269) Ó FEBS 2002 . The role of the second binding loop of the cysteine protease inhibitor, cystatin A (stefin A) , in stabilizing complexes with target proteases is exerted. total unitary free energy of binding of cystatin A to cathepsin B, demonstrating a major role of the second binding loop of cystatin A in the inhibition of this