Canonical protein inhibitors of serine proteases

CMLS, Cell. Mol. Life Sci. 60 (2003) 2427 – 2444 1420-682X/03/112427-18 DOI 10.1007/s00018-003-3120-x © Birkhäuser Verlag, Basel, 2003 CMLS Cellular and Molecular Life Sciences Review Canonical protein inhibitors of serine proteases D. Krowarsch, T. Cierpicki, F. Jelen and J. Otlewski* Institute of Biochemistry and Molecular Biology, University of Wroclaw, Tamka 2, 50-137 Wroclaw (Poland), Fax: + 48 71 375 26 08, e-mail: otlewski@protein.pl Received 28 March 2003; received after revision 12 May 2003; accepted 16 May 2003 Abstract. Serine proteases and their natural protein inhibitors are among the most intensively studied protein complexes. About 20 structurally diverse inhibitor families have been identified, comprising a-helical, b sheet, and a/b proteins, and different folds of small disulfiderich proteins. Three different types of inhibitors can be distinguished based on their mechanism of action: canonical (standard mechanism) and non-canonical inhibitors, and serpins. The canonical inhibitors bind to the enzyme through an exposed convex binding loop, which is complementary to the active site of the enzyme. The mecha- nism of inhibition in this group is always very similar and resembles that of an ideal substrate. The non-canonical inhibitors interact through their N-terminal segment. There are also extensive secondary interactions outside the active site, contributing significantly to the strength, speed, and specificity of recognition. Serpins, similarly to the canonical inhibitors, interact with their target proteases in a substrate-like manner; however, cleavage of a single peptide bond in the binding loop leads to dramatic structural changes. Key words. Serine protease; protein inhibitor; canonical conformation; protein-protein recognition. Introduction to protein inhibitors of proteases Proteases carry out an unlimited number of hydrolytic reactions both intra- and extracellularly [1]. They are found in viruses and in all living organisms from bacteria to mammals. Beside their physiological necessity, proteases are potentially hazardous to their proteinaceous environment and their activity must be precisely controlled by the respective cell or organism. When not properly controlled, proteases can be responsible for serious diseases. The basic level of control is normally achieved by regulated expression/secretion, by activation of proproteases [2], and by degradation of the mature enzymes. A second level of regulation is by inhibition of their proteolytic activity. Almost all known naturally occurring inhibitors directed toward endogenous cognate proteases are proteins; only some microorganisms se* Corresponding author. crete small non-proteinaceous compounds which block the host protease activity. Inhibition of proteases by proteins itself sounds paradoxical. Nevertheless, there are very common examples of inhibition of proteases by structurally unrelated proteins [3–5]. In fact, inhibitor structures, modes of inhibition and the nature of enzyme-inhibitor complexes are surprisingly different (table 1). In the past, inhibitors were believed to be specific for one of the four mechanistic classes of proteases (serine, cysteine, aspartic, or metalloproteases). While this is probably true in a prevailing number of cases, there are also known examples of proteins that are able to inhibit cysteine and aspartic protease [6], a serine and metalloprotease [7, 8], a serine and aspartic protease [9, 10], or a serine protease and amylase [11, 12], employing different, non-overlapping binding sites. The latest studies, compared to older ones, include testing of a broader range of proteases and some other hydrolases. However, very few examples of cross-protease 2428 D. Krowarsch et al. Protein inhibitors of serine proteases Table 1. Major features of protease inhibitors of serine, cysteine, metallo-, and aspartic proteases. Protease Inhibitor Examples Major features of inhibition Size Serine canonical inhibitors BPTI, OMTKY3, eglin 3–21 kDa per domain noncanonical inhibitors hirudin, TAP, ornithodorin tight, non-covalent interaction resembling enzyme-substrate Michaelis complex, direct blockage of the active site, no conformational changes, antiparallel b sheet between enzyme and inhibitor [3], similar mode of interaction through canonical protease-binding loop despite completely different inhibitor structures [84], important role of P1 residue, additive effects on association energy [147] extremely strong and specific interaction so far known for factor Xa and thrombin only, two-step kinetics, inhibition of the active site through N terminus of the inhibitor, two areas of interaction [150] irreversible covalent acyl-enzyme complex, huge conformational changes in inhibitor, disruption of protease active site [17, 151] serpins Cysteine cystatins thyropins IAP chicken cystatin, cystatin C, stefin B, kininogen p41, equistatin XIAP, cIAP1 CrmA, p35 Metallo PCI SMPI Pseudomonas aeruginosa inhibitor, Erwinia chrysanthemi inhibitor TIMP1, TIMP2, TIMP3, TIMP4 Aspartic IA3 PI-3 6–8 kDa per domain 45–55 kDa extremely tight but not specific, reversible non-covalent inhibition [26], interaction through two hairpin loops and N terminus forming a wedge, catalytic Cys25 accessible in complex, important interactions through P2 position [29] 11 – 13 kDa, up to 60 –120 kDa (kininogen) very tight inhibition, mechanism similar to cystatins but often more specific [27, 152] highly specific inhibition, reversible tight binding kinetics [153], inhibition also through an interdomain flexible linker region as non-productive binding in the opposite orientation to the substrates [30– 32] highly specific inhibition [34], similar to serpin mechanism of complexation [153] non-specific inhibition [34], irreversible acyl-enzyme, p35 N terminus shields catalytic Cys360 from water molecules, gross conformational changes in inhibitor [33] 7 kDa per domain 9 kDa per BIR domain tight enzyme-product complex, inhibition through C-terminal segment [36], key role of Val38 (P1) [154], no conformational changes in inhibitor upon complexation [155] rather specific inhibitor, inhibition mechanism resembling standard mechanism of canonical inhibitors of serine proteases, temporary inhibition [37], rigid protease-binding loop [37, 156, 157] both tight and weak inhibition observed, major interactions through five N-terminal residues, N-terminal amino group forms a coordinative bond to catalytic Zn, in analogy to TIMPs [40, 158] 4 kDa tight but not highly specific non-covalent interaction [159], N terminus and five inhibitor loops form wedge contacting the active site, bidental coordination of catalytic Zn through N terminus, major interactions through P1¢ residue, moderate conformational changes in inhibitor upon complexation [160] strong and highly specific [161], fully unfolded in free state, forms long helix in the complex comprising only Nterminal half of inhibitor, non-covalent complex [41]. strong but not highly specific [162], antiparallel b sheet formation between enzyme and inhibitor, no conformational changes [42] 38 kDa 35 kDa 11 kDa 15 kDa 20 – 22 kDa 8 kDa 17 kDa BPTI, bovine pancreatic trypsin inhibitor; OMTKY3, turkey ovomucoid third domain; TAP, tick anticoagulant peptide; IAP, inhibitor of apoptosis; XIAP, X-linked IAP; cIAP1, cellular IAP protein 1; BIR, baculoviral IAP repeat; CrmA, cytokine response modifier A; PCI, potato carboxypeptidase inhibitor; SMPI, Streptomyces proteinaceous metalloprotease inhibitor; TIMP, tissue inhibitors of metalloproteases; IA3, inhibitor of aspartic protease from Saccharomyces cerevisiae; PI-3, Ascaris suum pepsin inhibitor 3. CMLS, Cell. Mol. Life Sci. Vol. 60, 2003 class inhibition at the same protease recognition site have been reported, but their number is growing [13]. This review is devoted to canonical protein inhibitors of serine proteases. To put serine protease inhibitors in a broader context of possible inhibition modes, a brief outline of protein inhibitors of the four mechanistic classes (no protein inhibitors of threonine proteases are known) is presented below (table 1). From a structural point of view, blocking of the enzyme active site is almost always achieved by docking of an exposed structural element, such as a single loop or protein terminus, either independently or in combination of two or more such elements. Interestingly, antibodies, despite huge structural variability of their antigen-binding loops, cannot recognize the active site of proteases, as they only bind to flat or convex protein surfaces [14]. There are three distinct types of serine protease inhibitors (fig. 1). Up to the late 1980s, the majority of known protein inhibitors were those of the serine enzymes, either substrate-like-binding canonical inhibitors blocking the enzyme at the distorted Michaelis complex reaction stage [4, 15, 16], or serpins (serine protease inhibitors). While canonical inhibitors are small proteins, serpins are much larger, typically 350–500 amino acids in size, distributed from viruses to mammals [13, 17]. They are abundant in human plasma and mutations in serpins lead to numerous serious genetic diseases in humans [18]. Similarly to canonical inhibitors, serpins interact with their target proteases in a substrate-like manner. However, while the protease-binding loop of canonical inhibitors is kept in a wellordered conformation, the binding loop of serpins is much Review Article 2429 longer, about 17 amino acids, and able to adopt different conformations. In the active serpin, the binding loop protrudes significantly from the serpin scaffold [19], while in a much more stable latent conformation, this segment is inserted into the middle of the central b sheet A (fig. 1) [20]. In contrast to canonical inhibitors, serpins utilize the kinetic features of a hydrolytic reaction to form a very stable acyl-enzyme intermediate. The enzyme-serpin complex is a covalent acyl-enzyme adduct and upon acylation the protease is translocated by over 70 Å from its initial recognition site [21]. Serpins are the only family of serine protease inhibitors for which complex formation with non-serine enzymes – cysteine proteases [22] and aspartyl proteases [23] – has been demonstrated. Non-canonical inhibitors interact through their N-terminal segment which binds to the protease active site forming a short parallel b sheet. These inhibitors also form extensive secondary interactions with the target protease outside the active site, which provide additional buried area and contribute significantly to strength, speed, and specificity of recognition. The classic example is recognition of thrombin by hirudin [24]. Interestingly, such interactions are also formed by proteins possessing canonical-inhibitor-like folds or by Kazal-type inhibitors but with a distorted conformation of the binding loop. The non-canonical inhibitors are much less abundant than canonical inhibitors or serpins as they only occur in blood-sucking organisms and inhibit proteases involved in clot formation – thrombin or factor Xa. Only a few of them have been characterized in terms of structure and kinetics of interaction with the target protease. Figure 1. Examples of serine protease-inhibitor complexes: canonical, CMTI:trypsin (PDB: 1ppe); non-canonical, ornithodorin:thrombin (1toc); serpin, a-1 antitrypsin:trypsin (1ezx). The binding loop and P1 side chain residue of CMTI, a-1 antitrypsin and the N terminus of ornithodorin are marked in red. Secondary binding sites of ornithodorin are marked in orange. Secondary-structure elements are colored in blue (b sheets) and green (a helices). 2430 D. Krowarsch et al. The second largest and most carefully investigated group are inhibitors of cysteine proteases (table 1). As with the serine protease inhibitors, they can be divided into several groups: cystatins, stefins, kininogen, and thyroglobulin type 1 proteins [25, 26]. While the first three groups share similar structural features and form the cystatin superfamily, the latter group is structurally different [27]. Interestingly, all these proteins, and probably also other inhibitors [28], share clear similarities in the mode of their interaction with the enzymes. Members of the cystatin superfamily are reversible, extremely fast- and tight-binding competitive inhibitors, with the inhibition constants even in a low picomolar range [26]. The crystal structures of chicken cystatin and stefin B in complex with papain show a precise steric fit of a hydrophobic wedge-shaped edge of the cystatin inserted into the active site of the protease [29]. Inhibitors of caspases, the cysteine proteases responsible for apoptosis, are surprisingly variable, structurally unrelated to the cystatin superfamily, and much more specialized to fulfill their function. One group of caspase inhibitors, called inhibitors of apoptosis (IAPs), contains one or more baculoviral IAP repeat (BIR) domain that binds to the active site in a non-productive manner [30 –32]. Unexpectedly, a flexible linker connecting the BIR domains can also serve to block the active site of caspases 3 and 7. Another group of caspase inhibitors is the p35 family that has only been found in some baculoviruses. These inhibitors are able to inhibit different caspases using mechanism-based inactivation through formation of a covalent thiol ester [33]. Although the mechanism of inhibition in principle is similar to that of serpins, the molecular rearrangement of the inhibitor upon cleavage of the peptide bond is clearly different. Caspases can also be inhibited by viral serpins. The poxvirus CrmA serpin is able to block both caspases 1 and 8 and the serine protease granzyme B [34]. Although the structure of the CrmA-serpin complex remains to be determined, CrmA most probably undergoes an extensive structural transition during complex formation [35]. There are three unrelated groups of metalloprotease inhibitors, which inhibit their cognate enzymes through completely different mechanisms (table 1). The first discovered was potato carboxypeptidase A inhibitor (CPI), a small 39-residue protein, forming an enzyme-product type of complex (Gly39 is cleaved off in the crystal structure of the complex) through insertion of its C terminus into the active site of the enzyme [36]. Subsequently, a bacterial inhibitor, SMPI, was isolated from Streptomyces nigrescens that, surprisingly, appeared to inhibit metalloproteases of the gluzincin family by the standard mechanism of inhibition, extremely popular among canonical inhibitors of serine proteases [37]. Tissue inhibitors of metalloproteases (TIMPs) fold into a continuous wedge which blocks the active site of various matrix metalloproteases (MMPs) [38, 39]. This interaction is not highly Protein inhibitors of serine proteases specific and the mode of inhibition resembles the interaction between serralysins and their bacterial inhibitors from Pseudomonas aeruginosa and Erwinia chrysanthemi [40]. Although TIMPs and the two bacterial inhibitors are structurally unrelated, both groups of inhibitors form similar coordinative bonds to the catalytic Zn utilizing the N-terminal residue. Protein inhibitors of aspartic proteases are rare. Currently, structures of only two inhibitors have been reported (table 1). A small yeast protein IA3, composed of 68 amino acids, is able to inhibit aspartic protease A from the same organism in a very unusual way. While it shows no detectable secondary structure in solution, upon complexation with the enzyme, residues 2 –32 adopt an almost perfectly helical conformation revealing that the protease body serves as a folding template [41]. The PI-3 inhibitor from the intestinal parasite Ascaris suum presents an unrelated mode of inhibition. Its N-terminal b strand pairs with one strand of the active-site flap, forming an extensive eight-stranded b sheet spanning both proteins [42]. Canonical inhibitors The largest group of protein inhibitors are canonical inhibitors that act according to the standard mechanism of inhibition [15]. A huge number of canonical inhibitors have been described, isolated from various cells, tissues, and organisms; they often accumulate in high quantities especially in plant seeds, avian eggs, and various body fluids. Serine proteases and their natural protein inhibitors belong to the most intensively studied models of protein-protein recognition [14, 43]. Canonical inhibitors are widely distributed in essentially all groups of organisms and comprise proteins from 14 to about 200 amino acid residues. Canonical protein inhibitors do not form a single group but can be divided into different families. The segment responsible for protease inhibition, called the protease-binding loop, surprisingly has always a similar, canonical conformation in all known inhibitor structures [16, 44]. This convex, extended and solvent-exposed loop is highly complementary to the concave active site of the enzyme. The standard mechanism implies that inhibitors are peculiar protein substrates containing the reactive site P1-P1¢ peptide bond located in the most exposed region of the protease-binding loop (P1, P2 and P1¢, P2¢ specify inhibitor residues amino- and carboxy-terminal to the scissile peptide bond, respectively; S1, S2 and S1¢, S2¢ denote the corresponding subsites on the protease [45]). The reactive site can be selectively hydrolyzed by the enzyme, but the equilibrium value of this cleavage is often close to 1 at neutral pH, i. e., the reactive site can be cleaved to the extent of about 50 %. It is usually assumed, and very often verified experimentally, that standard- CMLS, Cell. Mol. Life Sci. Vol. 60, 2003 mechanism inhibitors possess canonical conformation of the binding loop. On the other hand, while loops of canonical conformations occasionally occur in various proteins, there is no evidence that they can block proteases according to the standard mechanism [44, 46]. Classification A classification of canonical inhibitors was originally proposed by Laskowski and Kato in 1980 [15]. At that time they could distinguish eight families, based mainly on the disulfide bond topography, location of the reactive site, and sequence homology. Currently, 18 inhibitor families are recognized [5] (table 2). Crystal and/or solution structures are known for representatives of almost all families (table 2). Since the inhibitors are small, rigid, and stable, these structures have often been determined with high resolution and accuracy. Furthermore, serine proteases and enzyme-inhibitor complexes crystallize easily, often providing high-resolution data. Extensive structural information is also available for protease-inhibitor complexes. The global structures of proteins representing different inhibitor families are completely different. Most often they comprise either purely b sheet or mixed a/b proteins; they can also be a-helical or irregular proteins rich in disulfide cross-links. It is often stressed that the canonical inhibitors represent the most distinct and extensive example of convergent protein evolution, since a similar function has been implemented many times during evolution through preservation of the canonical loop conformation in many unrelated proteins [5]. Examples of different folds of inhibitor structures are shown in figure 2. The inhibitor structure The inhibitor scaffolds are of very different structural types. In several inhibitor families like BPTI, Kazal, potato 1 and 2, cereal, SSI, STI, and ecotin, typical secondary-structure elements together with the presence of a hydrophobic core are found. In other families, including squash, Bowman-Birk, grasshopper, hirustasin, chelonianin, and Ascaris there is essentially a lack of both hydrophobic core and extensive secondary structure. For these inhibitors, disulfide bonds, which are usually buried inside the molecule, are the major determinant of protein stability and/or rigidity. With respect to the cross-links, the most unusual is probably a cyclic peptide from sunflower seeds that strongly inhibits trypsin (SFTI-1). This cyclic inhibitor is homologous to the Bowman-Birk family and consists of two antiparallel b strands additionally crosslinked with a disulfde bond and stabilized by numerous hydrogen bonds [47]. Natural cyclic peptides also occur Review Article 2431 among squash inhibitors [48]. Worth mentioning is that it is possible to synthesize chemically a cyclic version of the non-cyclic protease inhibitor BPTI [49, 50]. The presence of an inhibitory domain is usually indicative of serine protease inhibition. However, examples are also known of naturally occurring inhibitors belonging to ovomucoids, e.g., with Pro at P1, that are very poor inhibitors of all proteases tested including prolyl endopeptidase [51]. Sometimes other functions, not related to canonical inhibition, could be detected for canonical domains. For example, Kazal domains occur in follistatin [52], the squash inhibitor fold is found in the metalloprotease inhibitor PCI [53], and the BPTI domain is frequently observed in snake potassium channel blockers or linked to the phospholipase A2 domain, Alzheimer precursor protein [54], or collagen a1 and a3 chains [55]. In one case tested of the non-inhibitory C5 domain of collagen VI, a strong protease inhibitor could be generated through multiple substitutions in the binding-loop region [56]. However, in many other cases, conversion of a noninhibitory to inhibitory protein required more effort due to severe conformational and dynamic changes in the binding-loop region. Two non-canonical inhibitors of coagulation proteases from the soft tick, ornithodorin and tick anticoagulant peptide (TAP), surprisingly show a scaffold of the archetypical canonical inhibitor BPTI (fig. 1). Ornithodorin contains two BPTI-like domains containing insertion/ deletion in the binding-loop segments, which lead to their major distortion [57]. In fact, this binding loop does not contact the protease, but as in the hirudin-thrombin complex, the N-terminal tail of ornithodorin penetrates the thrombin active site and forms a parallel b sheet with the thrombin Ser214-Gly219 segment. Similarly, TAP, which is a strong inhibitor of factor Xa [58], interacts through the N terminus with the active site of factor Xa [59]. Inhibitors belonging to different families are generally stable or even extremely stable proteins with high denaturation temperatures and resistance to chemical denaturants. At neutral pH, BPTI shows a denaturation temperature (Tden) of about 100 °C and is stable in 6 M guanidinium chloride [60, 61]. Kazal inhibitors denature with a Tden up to 90°C and are also stable in 6 M GdmCl [62, 63], and STI unfolds at 65 °C [64]. For small inhibitors like those from the squash family, cooperative denaturation could not be demonstrated due to a lack of secondary structure, hydrophobic core and too small size of the cooperative unit. Inhibitors are often heavily cross-linked with conserved disulfide bonds. The topology of the disulfide bonds is usually well preserved within a single family. However, some members of the potato 1 family show either a single disulfide [CMTI-V (Cucurbita maxima trypsin inhibitorV) and LUTI (Linum usitatissimum trypsin inhibitor)] or no disulfide (CI-2 and eglin c). Within the STI family 2432 D. Krowarsch et al. Protein inhibitors of serine proteases Table 2. Representative X-ray and NMR three-dimensional structures of protein inhibitor families of serine proteases and their enzyme complexes. Family Total number Free inhibitor Enzyme-inhibitor complex number representative PDB resolution Å number representative PDB resolution Å BPTI: rat trypsin OMTKY: chymotrypsin eglin c: subtilisin CPTI II: trypsin ecotin: crab collagenase STI: porcine trypsin MbBBI: Ns3protease SFTI-1: trypsin bdellastasin: porcine trypsin C/E-1 inhibitor: porcine elastase PMP-C: chymotrypsin SSI: subtilisin BPN¢ PCI 1: SGPB 1f7z 1.5 1cho 1.8 1cse 1.2 2btc 1.5 1aaz 2.3 1avw 1.75 1df9 2.1 1sfi 1.6 1eja 2.7 1eai 2.4 1gl1 2.1 2sic 1.8 4sgb 2.1 BPTI 63 28 BPTI 5pti 1.0 35 Kazal 36 16 OMSVP3 2ovo 1.5 20 Potato 1 27 16 CI-2 2ci2 2.0 11 Squash 17 9 CMTI I 1lu0 1.03 8 Ecotin 12 3 ecotin 1ifg 2.0 9 STI 11 9 STI 1avu 2.3 2 BBI 9 6 BBBI 1c2a 1.9 3 BBI (SFTI) 2 1 SFTI-1 1jbl NMR 1 Antistasin 7 3 hirustasin 1bx7 1.2 4 Ascaris 7 6 AMCI 1ccv NMR 1 Grasshopper 6 4 PMP-C 1pmc NMR 2 SSI 5 1 SSI 3ssi 2.3 4 Potato 2 4 3 T1 1tih NMR 1 Cereal Chelonianin 4 2 4 1 CHFI R-elafin 1bea 2rel 1.95 NMR 1 Rapeseed 1 1 ATTp 1jxc NMR Elafin: porcine elastase 1fle 1.9 Arrowhead In the absence of X-ray structure, the representative NMR structure is indicated. Total number, total number of structures deposited in PDB (Protein Data Bank) (free structures and complexed with protease); OMSVP3, silver pheasant ovomucoid third domain; OMTKY, turkey ovomucoid; CI-2, chymotrypsin inhibitor 2; CMTI I, Cucurbita maxima trypsin inhibitor I; CPTI II, Cucurbita pepo trypsin inhibitor II; STI, soybean trypsin inhibitor; BBBI, barley Bowman-Birk inhibitor; SFTI-1, sunflower trypsin inhibitor; MbBBI, mung bean BowmanBirk inhibitor; AMCI, Apis mellifera chymotrypsin inhibitor; C/E-1 inhibitor, Ascaris chymotrypsin/elastase inhibitor 1; PMP-C, Pars intercerebralis major peptide; SSI, Streptomyces subtilisin inhibitor; T1, trypsin inhibitor from Nicotiana alata; SGPB Streptomyces griseus protease B; CHFI, corn Hageman factor inhibitor; ATTp, Arabidopsis thaliana trypsin inhibitor. also, there is an inhibitor with a single disulfide bond instead of the two typically observed among members of this family [65]. Worthy of note is that engineering of an additional disulfide bond near the reactive site of the silver pheasant third domain (Kazal inhibitor) left intact its potent inhibitory activity toward chymotrypsin, Streptomyces griseus proteases A and B, but almost abolished it toward pancreatic elastase [66]. Selective reduction or elimination of disulfide bond(s) in inhibitors belonging to different families usually leads to a significant destabilization of the inhibitor molecule, to a lower association-energy and larger sensitivity to proteolysis [67–70]. The same holds for destabilizing mutation(s) introduced into the inhibitor core [71, 72]. CMLS, Cell. Mol. Life Sci. Vol. 60, 2003 Review Article 2433 Figure 2. Representative set of canonical inhibitor structures. The following structures are shown: hirustasin (PDB: 1bx7); AMCI (1ccv); PI-II, protease inhibitor-II (1pi2); STFI-1 (1jbl); CHFI (1bea); R-elafin (2rel); ecotin (1ifg); PMP-C (1pmc); OMSVP3 (2ovo); BPTI (5pti); STI (1avu); CI-2 (2ci2); T1 (1tih); ATTp (1jxc); SSI (3ssi); CMTI I (1lu0). Binding loop and P1 side chain residue are marked in red, secondary-structure elements are colored in blue (b sheets) and green (a helices). Name of inhibitor family (in bold) and name of representative are indicated. Domain architecture Many inhibitors are single-domain proteins. With a possible exception for the arrowhead protease inhibitor [73], the single inhibitory domain contains only one reactive site responsible for protease binding [74]. This is true for all members of the arrowhead, Ascaris, ecotin, STI, potato 1, cereal, rapeseed, silkworm, SSI, and squash families. In the remaining families (antistasin, BowmanBirk, BPTI, grasshopper, Kazal, chelonianin, and potato 2) the single domain can be repeated 2, 3, 4, 5, 6, 7, 9 or even 15 times to form a multidomain, single-chain inhibitor which is able to interact independently with sev- eral protease molecules at their reactive sites belonging to separate domains. The different organization observed in multidomain proteins is shown in figure 3. In the simplest case, represented by ovomucoids, three homologous Kazal domains, each cross-linked with three disulfide bonds, are connected by short and flexible linkers. These domains are independently able to inhibit serine proteases – for example the first domain of turkey ovomucoid inhibits Gluspecific S. griseus protease, the second domain inhibits trypsin and the third blocks chymotrypsin, elastase and subtilisin. Similarly, the first domain of tissue factor pathway inhibitor (TFPI) composed of three tandemly 2434 D. Krowarsch et al. arranged BPTI domains inhibits factor VIIa/tissue factor, the second interacts with factor Xa, and the third is without detectable inhibitory function [75]. In contrast, in the crystal structure of bikunin, composed of two BPTI domains, the binding loop of the second domain is obstructed by the first domain, thus affecting protease binding (fig. 3) [76]. A highly unique domain topology occurs in a protein inhibitor precursor from Nicotiana alata (NaProPI). This protein is composed of six homologous repeats; however, their sequences do not coincide with the structural domains [77]. One of these domains comprises two chain fragments from the first and last repeat, strongly suggesting that the precursor adopts a circular bracelet-like structure (fig. 3). In Bowman-Birk inhibitors, there are again two independent binding loops but the presence of seven inter- and intradomain disulfides results in a much more compact two-domain protein. Two Kazal domains of rhodniin serve to inhibit one enzyme – thrombin. While the N-terminal domain binds to the active site of thrombin through its canonical loop (interestingly with His at P1, despite strong preference of thrombin for Arg), the binding loop of the C-terminal domain is distorted, thus excluding canonical interaction [78]. Instead, this domain recognizes the fibrinogen exosite on the thrombin surface using residues located outside the binding loop region (fig. 3). A homodimeric inhibitor from Escherichia coli, called ecotin, although built of a single domain is active as a dimer in which both monomers provide the proteasebinding surface (fig. 3) [79]. A comparison of the inhibitory properties of dimeric ecotin with that of an engineered monomeric form reveals a complex and non-additive interplay of binding energies provided by the binding sites located on the two subunits [80]. Finally, several multidomain proteins containing combinations of WAP (whey acidic protein), Kunitz, Kazal, and thyroglobulin domains have been identified that may control multiple types of serine, aspartic, metallo and cysteine proteases [81]. Protein inhibitors of serine proteases Figure 3. Examples of multidomain serine protease inhibitors: model of Na-Pi structure based on C1T1 and C2 NMR structures (PDB: 1fyb and 1qh2); ecotin dimer (1ecz); model of ovomucoid structure based on OMSVP3 crystal structure (2ovo); BBI (1bbi); bikunin (1bik); rhodniin (1tbq). Binding loop and P1 side chain residue are marked in red, residues involved in protease dimerization by ecotin are marked in yellow, residues from the C-terminal domain of rhodniin interacting with thrombin are marked in cyan. Name of inhibitor family (in bold) and name of presented structures are indicated. The canonical conformation of the binding loop The convex protease-binding loop exhibits an extended conformation, which significantly protrudes from the protein scaffold and serves as a rather simple recognition motif (fig. 4). The loop forms a sequential epitope spanning positions P3 to P3¢. Residues that precede or follow this segment (e.g., P6-P4 or P4¢) and residues from a sequentially remote region, called the secondary contact region, can also contact the enzyme and influence the association energy [82 –84]. The central section of the loop contains a solvent-exposed P1-P1¢ peptide bond, called the reactive site. This bond is not fully inert to the proteolytic attack by the cognate enzyme – the equilibrium value of Figure 4. Superimposition of main chain P3-P3¢ segments (according to Schechter and Berger notation) of binding loops of: OMSVP3, light gray (PDB: 2ovo); SSI, gray (3ssi); BPTI, dark gray (PDB: 5pti). CMLS, Cell. Mol. Life Sci. Review Article Vol. 60, 2003 the reactive-site peptide bond opening, called the hydrolysis constant, is usually not far away from unity [85 –87]. The conformation of the cleaved inhibitor is very similar to the intact form with clear exceptions for the local structural changes near the P1-P1¢ peptide bond [88, 89] and increased internal mobility of the cleaved loop, but not of the inhibitor scaffold [90, 91]. Thermodynamic analysis reveals that hydrolysis of the reactive site in the native inhibitor does not lead to a significant increase in entropy [92]. The full entropy gain is realized upon denaturation of the reactive-site-cleaved inhibitor which leads to predicted values of Khyd for the hydrolysis of the reactive site in the denatured inhibitor at between 100 and 1000 [64, 93]. However, in the case of CMTI-V, a large entropy increase was observed upon reactive-site cleavage of the native inhibitor [94]. The main chain conformations of the binding loops of free inhibitors representing different families are similar and become even more alike after complex formation with the enzyme (fig. 4) [44]. The canonical conformation is also presumed to be adopted by a productively bound protein substrate. The binding loops within one family (most intensively studied for ovomucoids belonging to Kazal inhibitors [84, 95]) often show a high level of sequential variability; nevertheless, in all cases studied, the loops had canonical conformation. A lack of hypervariability has been indicated for squash [96] and potato 1 families [97]. The amino acid sequences of the binding loops show many clear amino acid preferences in different families. For example, half cystine is present either at P3 (the Kazal, Bowman-Birk, grasshopper, silkworm, squash, SSI, potato 2, and Ascaris families) or at P2 (the BPTI, antistasin, arrowhead, hirustasin, and chelonianin families). Thr is often observed at P2 (the Kazal, potato 1, BowmanBirk, SSI, ecotin, and Ascaris families) and Pro is conserved at P3 in the STI family [44]. A conserved Ile is always present at the P1¢ position in squash inhibitors. Its mutation to Leu leads to a severe disorder of the binding loop [98]. A similar disorder of the loop was observed upon Asp46ÆSer substitution at the P1¢ site of eglin c [99]. Ala and Gly are highly conserved at P1¢ in the BPTI family. Introduction of larger side chains at this position leads to a huge decrease in the association constants with proteases [100]. In Bowman-Birk inhibitors, prolines are frequently observed at P3¢ and P4¢. The Pro at P3¢ in geometry is required for strong inhibition while that at P4¢ stabilizes the P3¢ configuration [101]. Thus, the canonical conformation may be achieved by many unrelated sequences. The canonical loop conformation results from a rather extensive system of disulfide bond(s), hydrogen bonds, and/or hydrophobic interactions, which involve residues both from the loop and the inhibitor scaffold. A scheme 2435 of the loop-maintaining interactions in different families is shown in figure 5. For example, in the OMSVP3 (Kazal family) inhibitor, the carbonyl oxygens of P2 and P1¢ are involved in hydrogen bonds to the side chain of Asn33, and in CI-2 (potato 1 family), these carbonyls are hydrogen bonded to two arginines at P6¢ and P8¢. The LUTI inhibitor which belongs to the same family, despite a Trp Æ Arg substitution at P8¢, is a strong inhibitor of trypsin and its solution structure shows that loop conformation is well preserved [102]. Interestingly, mutation of Arg at P8¢ to Lys in the homologous eglin c led to a destabilization of the binding loop [99]. There is evidence based on changes in 15N relaxation rates for an increased dynamics of the loop in CMTI-V mutants with eliminated side chains of Arg50 or Arg52 [103, 104]. In many inhibitor families, there is also a common hydrogen bond between the side chain or main chain of the P2 and P1¢ positions (fig. 5). In ovomucoid third domains, the side chain-side chain hydrogen bond between Thr17 (P2) and Glu19 (P1¢) is long in the free inhibitor (and therefore not shown on fig. 5) but becomes shortened by about 0.5 Å upon complex formation, thus showing that it energetically favors the complex [105]. In Ascaris trypsin inhibitor (ATI), there is a pH-induced structural transition in the binding-loop region observed by nuclear magnetic resonance (NMR) – while the loop is rigid and canonical at pH 2.4, it becomes disordered at pH 4.75 [106]. This conformational change likely results from deprotonation of Glu32 and disruption of the hydrogen bond between side chains of Thr30 (P2) and Glu32 (P1¢), since in a homologous inhibitor, AMCI, which has Gln instead of Glu, no pH-induced structural transition is observed [107]. Interestingly, replacing the CI-2 loop sequence with that of helix E from subtilisin Carlsberg led to a protein hybrid with a well-preserved scaffold and extended loop-like conformation of the introduced sequence [108]. This result shows that the context of the CI-2 scaffold is sufficient to impose appropriate loop conformation. Furthermore, a multiple mutant of BPTI with almost all residues in the binding loop replaced with alanines was constructed. Although this loop region was significantly less structured compared to the wildtype protein, the binding loop still could adopt the proper conformation and interact with trypsin and chymotrypsin [109]. The inhibitor scaffold, therefore, seems to play an active role in maintaining the loop conformation. The standard mechanism The canonical inhibitor-cognate protease interaction is preserved in all cases tested and called the standard mechanism [15]. The interaction between enzyme and inhibitor can be presented in a simplified form as a hydro- 2436 D. Krowarsch et al. Protein inhibitors of serine proteases Figure 5. Binding-loop structures of representatives of inhibitor families: bdellastasin (PDB :1c9p); AMCI (1ccv); BBBI (1c2a); CHFI (1bea); elafin (1fle); ecotin (1ifg); PMP-C (1gl1); OMJPQ3, Japanese quail ovomucoid third domain (1ovo); BPTI (5pti); STI (1ba7); eglin c (1cse); PCI-1 (4sgb); CMTI I (2sta); SSI (3ssi). The fragments comprise P3-P2¢ segments with additional residues involved in binding-loop stabilization. The backbone is shown in gray, P1 side chains in red, disulfide bridges in yellow, polar and hydrophobic side chains in green and magenta, respectively. Additionally, hydrogen bonds are shown as black dashed lines. Amino acid residues are labeled using one-letter codes including Schechter and Berger notations. Name of inhibitor family (in bold) and name of representative are indicated. CMLS, Cell. Mol. Life Sci. Review Article Vol. 60, 2003 lysis/resynthesis reaction of the P1-P1¢ reactive-site peptide bond: kon E+I ¤ koff k*off EI ¤ E + I* k*on (1) where E is the protease, I is the inhibitor, I* is the reactive-site-cleaved inhibitor, EI is the stable complex, kon and k*on are respective second-order association rate constants, and koff and k*off are respective first-order dissociation rate constants of the complex. A recently described bacterial inhibitor of metalloproteases appears to resemble canonical inhibitors with respect to the inhibition mechanism [110]. Compared to peptide bond hydrolysis in a regular protein: 1) The complex EI is much more stable than the Michaelis ES complex. Typical inhibition constant (Ki) values are 106- to 109-fold lower than Km values. Often, complex can be easily crystallized and shows all typical features of protein-protein recognition [14, 43]. 2) The catalytic rate constant for the hydrolysis of the reactive site is extremely low at neutral pH [87, 111]. However, there are examples of hydrolysis of reactive sites by inhibited proteases, proceeding at much higher rates [112, 113]. 3) The conserved mode of recognition between the protease-binding loop and the active site leads to many different serine proteases (belonging both to the chymotrypsin and subtilisin families) of different specificities being inhibited at the same reactive site in the case of the turkey ovomucoid third domain [74]. This is also true for other inhibitors. Eglin c (potato 1 family) inhibits 14 serine proteases with association constant values greater than 108 M–1 [5]. Furthermore, the three-dimensional structures of BPTI complexed with trypsin [114], chymotrypsin [115], pancreatic kallikrein [116], thrombin [117], factor VIIa [118], and trypsinogen [119] show extremely similar modes of recognition, despite one billion-fold difference in their affinities. A similar difference in the association constant exists for the interaction of trypsin with ten P1 mutants of BPTI and, again, crystal structures of the respective complexes show virtually identical modes of recognition [120]. 4) The kcat/Km index for the hydrolysis of the reactive-site peptide bond is often about 106 M–1 s–1, suggesting that inhibitors are good substrates [121]. However, this parameter describes the enzyme-substrate reaction only at low substrate concentrations ([S]<Km). Since the Km values for hydrolysis are extremely low, the reaction rate is proportional to kcat which is known to be extremely low in the case of reactive-site hydrolysis at neutral pH. 5) The hydrolysis reaction is reversible, i.e. the cleaved inhibitor is active and forms the same complex with 2437 the enzyme as the intact form. During complex formation, resynthesis of the reactive-site peptide bond occurs [111]. The kinetic parameters for reactive-site resynthesis are often similar to those of the hydrolysis reaction. The phenomenon of hydrolysis/resynthesis also occurs at other peptide bonds of the binding loop and can also be catalyzed by non-serine proteases [87]. 6) The equilibrium value of [I*]/[I] (hydrolysis constant, Khyd) is often close to unity (i. e., about 50% of the inhibitor molecules have the reactive site cleaved) at pH 6 where Khyd is pH independent [86, 87, 121], but examples are known of natural ovomucoid third domain variants with Khyd in the range of 0.4–35 [85]. Typical Khyd values for a single peptide bond hydrolysis in a native protein containing secondary structure are in the neglibly low values of 10–3 to 10–8 [122]. Thus the values of Khyd in protein inhibitors are extremely high. Substitution of the residues which maintain the binding loop conformation affects the value of Khyd. However, the effect of a single mutation is relatively small and does not exceed the factor 3 –5 [85]. 7) While the kon values for protease-inhibitor association are typically about 106 M–1 s–1, the koff values may differ by many orders of magnitude. The k*on values can also differ by many orders of magnitude for the interaction of an inhibitor with different proteases [74]. 8) At high concentrations of enzyme and inhibitor, the existence of additional unstable loose complexes L and L* can be detected by stopped-flow methods [121]. Recent experiments revealed that the acyl-enzyme intermediate could be formed rapidly, suggesting that there is no energetic barrier to the acylation reaction [123]. The protease-inhibitor complex The mode of recognition between different canonical inhibitors and serine proteases is always almost the same. In the stable complex, which was a subject of numerous crystallographic studies, a short antiparallel b sheet is formed between the P3-P1 residues and the 214 – 216 (Ser125-Gly127 in subtilisin) segment of the enzyme (fig. 6). The energetic contribution of one of these intermolecular main chain hydrogen bonds (donated by the HN amide of the P1 residue of OMTKY3) was recently found to be about 1.5 kcal/mol [124]. There is an additional antiparallel b sheet between the P4-P6 fragment and Tyr104-Gly102 residues in subtilisin complexes, which does not exist in chymotrypsin-like enzymes [125, 126]. Other very important features of the complex include: a short (usually about 2.7 Å) contact between the P1 carbonyl carbon and the catalytic serine residue (significantly shorter in rhodniin-thrombin [78] and mung bean trypsin inhibitor-trypsin complexes 2438 D. Krowarsch et al. Protein inhibitors of serine proteases Figure 6. Schematic representation of canonical inhibition based on the structure of the CMTI I:trypsin complex (PDB: 1ppe). The inhibitor [dark gray, residues marked as (I)] binds to the protease [light gray, residues marked as (E)] in a manner similar to that of a typical substrate. Several characteristic interactions are shown: (i) antiparallel b sheet formed between residues P1-P3 of the inhibitor and residues 214 –216 of the protease; (ii) sub-van der Waals contacts between Ser195 Og and P1 carbonyl carbon, and (iii) hydrogen bonds from the oxyanion binding hole (HN of Gly193 and Ser195) to the P1 carbonyl oxygen. Nitrogen and oxygen atoms are shown as dark and light gray balls, respectively. Amino acid residues are labeled using three-letter codes including Schechter and Berger notation. [127]), and two hydrogen bonds between the carbonyl oxygen of P1 and Gly193/Ser195 amides of the oxyanion binding hole, and the hydrogen bonds between the P1 HN group and the side chain of Ser195 and the carbonyl of Ser214. Conversion of the P2-P1 amide bond to an ester bond reduces the association free energy by about 1.5 kcal/mol [124, 128]. The reactive-site peptide bond remains intact in all crystallographically studied complexes. All the above-mentioned hydrogen bonds and the shape complementarity of interacting areas ensure very similar recognition modes between different proteases and inhibitors. In the complex, about 10 –18 amino acid residues of the inhibitor and 17 –30 residues of the protease make numerous interactions – mainly van der Waals (typically more than 100) and hydrogen bonds (about 8–15). The total area of the two components buried in the interface is about 1400 Å2. According to NMR relaxation studies, the protease-binding loop, which is often poorely structured in free inhibitors [90, 91, 129], becomes significantly rigidified in the complex. There are no significant conformational changes on either the enzyme or inhibitor part accompanying complex formation, with the exception of zymogen complexes. In the trypsinogen-inhibitor complex, major structural rearrangements are observed in the activation domain comprising the active-site region [130]. The organization of the activation domain in the complex with the inhibitor is remarkably similar to one observed in the active enzyme, but fully disordered in the free zymogen. The association constant is about 107-fold lower than for the active enzyme [131, 132]. Despite the inherently low activity of the zymogen, the standard mechanism works also for trypsinogen, which is able to resynthesize the reactive-site peptide bond of the inhibitor [133]. The P1 position In canonical inhibitors, position P1 determines to a large extent the protease-inhibitor association energy. With the exception of Trp, Ile and Cys, all amino acids have been observed at this position in inhibitors representing different families [134]. The plot of the substrate transitionstate energy log(kcat/Km) versus the enzyme-inhibitor association energy log(Ka) determined for a set of P1 oligopeptide substrates and protein inhibitors is a straight line with a slope not far from unity, suggesting that interactions within the S1 pocket do not change as the reaction proceeds from the enzyme-inhibitor complex to the transition state [51, 135, 136]. P1 Gly and particularly P1 Pro are very disfavored for binding with most of the proteases tested [51, 137]. Also the charged P1 side chains of Asp, Glu and His (but not their uncharged forms), when placed in hydrophobic S1 pockets, strongly oppose complex formation [138]. The shifts of the pK values of these side chains placed in the hydrophobic S1 pocket of SGPB reach 5 pH units. The P1 side chain is fully exposed in all free inhibitor structures (fig. 7) and becomes imbedded in the S1 pocket CMLS, Cell. Mol. Life Sci. Vol. 60, 2003 Review Article 2439 energies of protease-inhibitor interactions were extensively tested in the Laskowski laboratory for the interaction between ovomucoid third domains and six different serine proteases [147]. Depending on the experimental error analysis, the protease tested, and the number of mutations introduced, the additivity holds for from 60 to almost 100% of the analyzed cases. Additivity offers a superb possibility of creating strong (with Ka values up to 1017 M–1) or specific inhibitors for different proteases through careful design of multiple mutants not only of ovomucoid third domains but also of inhibitors belonging to all other families [147–149]. Very recently, additivity was successfully tested in the interaction between chymotrypsin and alanine-shaved mutants of BPTI [83]. Figure 7. Solvent accessible area of CMTI I (PDB: 1lu0). Only the main chain of the inhibitor and side chain of P1 (Arg5) are shown. The solvent-exposed protease-binding loop (residues P3 to P3¢) and corresponding surface is colored dark gray. Amino acid residues are labeled using Schechter and Berger notation. upon complex formation. It can form about 50% of the interface contact area and provide even up to 70 % of the association energy as deduced from comparisons with the P1 Gly variant [51, 137, 139]. Cognate P1 side chains enter the S1 pocket preserving optimal c angles [140, 141]. An improperly matched P1-S1 interaction in terms of size, shape, charge, polarity, or branching of the P1 side chain leads to severe effects on the association energy [135, 137, 139, 141, 142]. Furthermore, alanine-scanning mutagenesis of BPTI [82] and theoretical calculations of the protease-inhibitor interaction [143] clearly reveal a dominant role for the P1 residue. Since the P1 residue occupies a central part of the canonical loop, its substitutions in different inhibitor families often cause very similar energetic effects on the binding to the serine protease, a phenomenon called the interscaffolding additivity [137, 139]. A lack of interscaffolding additivity was observed for the interaction of P1 Lys variants of OMTKY3 and BPTI with chymotrypsin, and was explained based on the crystal structures of the two complexes in which completely different conformation of the P1 side chain and its interactions were observed upon binding in the respective S1 pockets [115, 137, 144]. The additivity In the simplest case, additivity assumes that the energetic effects of two individual mutations sum up, within experimental error, in a mutant containing the two mutations [145]. Additivity can also be applied in systems containing a larger number of mutations. In protein chemistry, additivity is tested in various macromolecular processes almost exclusively at the level of free energy effects [146]. Free Acknowledgements. The research of J. Otlewski was supported by an International Research Scholar award from the Howard Hughes Medical Institute and by a scholarship from the Foundation for Polish Science. 1 Neurath H. (1989) Proteolytic processing and physiological regulation. Trends Biochem. Sci. 14: 268 – 271 2 Khan A. R. and James M. N. (1998) Molecular mechanisms for the conversion of zymogens to active proteolytic enzymes. Protein Sci. 7: 815 – 836 3 Bode W. and Huber R. (2000) Structural basis of the endoproteinase-protein inhibitor interaction. Biochim. Biophys. Acta 1477: 241 – 252 4 Otlewski J., Krowarsch D. and Apostoluk W. (1999) Protein inhibitors of serine proteinases. Acta Biochim. Pol. 46: 531 – 565 5 Laskowski M. Jr, Qasim M. A. and Lu S. M. (2000) Interaction of standard mechanism, canonical protein inhibitors with serine proteinases. In: protein-Protein Recognition, pp. 228–279, Kleanthous C. (ed.), Oxford University Press, Oxford 6 Strukelj B., Lenarcic B., Gruden K., Pungercar J., Rogelj B., Turk V. et al. (2000) Equistatin, a protease inhibitor from the sea anemone Actinia equina, is composed of three structural and functional domains. Biochem. Biophys. Res. Commun. 269: 732 – 736 7 Kumazaki T., Kajiwara K., Kojima S., Miura K. and Ishii S. (1993) Interaction of Streptomyces subtilisin inhibitor (SSI) with Streptomyces griseus metallo-endopeptidase II (SGMP II). J. Biochem. (Tokyo) 114: 570 – 575 8 Hiraga K., Suzuki T. and Oda K. (2000) A novel doubleheaded proteinaceous inhibitor for metalloproteinase and serine proteinase. J. Biol. Chem. 275: 25173 – 25179 9 Mares M., Meloun B., Pavlik M., Kostka V. and Baudys M. (1989) Primary structure of cathepsin D inhibitor from potatoes and its structure relationship to soybean trypsin inhibitor family. FEBS Lett. 251: 94 – 98 10 Ritonja A., Krizaj I., Mesko P., Kopitar M., Lucovnik P., Strukelj B. et al. (1990) The amino acid sequence of a novel inhibitor of cathepsin D from potato. FEBS Lett. 267: 13 – 15 11 Zemke K. J., Muller-Fahrnow A., Jany K. D., Pal G. P. and Saenger W. (1991) The three-dimensional structure of the bifunctional proteinase K/alpha-amylase inhibitor from wheat (PK13) at 2.5 Å resolution. FEBS Lett. 279: 240 – 242 12 Vallee F., Kadziola A., Bourne Y., Juy M., Rodenburg K. W., Svensson B. et al. (1998) Barley alpha-amylase bound to its endogenous protein inhibitor BASI: crystal structure of the complex at 1.9 Å resolution. Structure 6: 649 – 659 13 Gettins P. G. (2002) Serpin structure, mechanism, and function. Chem. Rev. 102: 4751 – 4804 2440 D. Krowarsch et al. 14 Jones S. and Thornton J. M. (1996) Principles of protein-protein interactions. Proc. Natl. Acad. Sci. USA 93: 13 – 20 15 Laskowski M. Jr and Kato I. (1980) Protein inhibitors of proteinases. Annu. Rev. Biochem. 49: 593–626 16 Bode W. and Huber R. (1992) Natural protein proteinase inhibitors and their interaction with proteinases. Eur. J. Biochem. 204: 433–451 17 Silverman G. A., Bird P. I., Carrell R. W., Church F. C., Coughlin P. B., Gettins P. G. et al. (2001) The serpins are an expanding superfamily of structurally similar but functionally diverse proteins: evolution, mechanism of inhibition, novel functions, and a revised nomenclature. J. Biol. Chem. 276: 33293–33296 18 Stein P. E. and Carrell R. W. (1995) What do dysfunctional serpins tell us about molecular mobility and disease? Nat. Struct. Biol. 2: 96–113 19 Elliott P. R., Abrahams J. P. and Lomas D. A. (1998) Wild-type alpha 1-antitrypsin is in the canonical inhibitory conformation. J. Mol. Biol. 275: 419–425 20 Tucker H. M., Mottonen J., Goldsmith E. J. and Gerard R. D. (1995) Engineering of plasminogen activator inhibitor-1 to reduce the rate of latency transition. Nat. Struct. Biol. 2: 442–445 21 Huntington J. A., Read R. J. and Carrell R. W. (2000) Structure of a serpin-protease complex shows inhibition by deformation. Nature 407: 923–926 22 Komiyama T., Ray C. A., Pickup D. J., Howard A. D., Thornberry N. A., Peterson E. P. et al. (1994) Inhibition of interleukin-1 beta converting enzyme by the cowpox virus serpin CrmA: an example of cross-class inhibition. J. Biol. Chem. 269: 19331–19337 23 Mathialagan N. and Hansen T. R. (1996) Pepsin-inhibitory activity of the uterine serpins. Proc. Natl. Acad. Sci. USA 93: 13653–13658 24 Stubbs M. T., Huber R. and Bode W. (1995) Crystal structures of factor Xa specific inhibitors in complex with trypsin: structural grounds for inhibition of factor Xa and selectivity against thrombin. FEBS Lett. 375: 103–107 25 Turk D., Sturzebecher J. and Bode W. (1991) Geometry of binding of the N alpha-tosylated piperidides of m-amidino-, p-amidino- and p-guanidino phenylalanine to thrombin and trypsin: X-ray crystal structures of their trypsin complexes and modeling of their thrombin complexes. FEBS Lett. 287: 133–138 26 Turk B., Turk V. and Turk D. (1997) Structural and functional aspects of papain-like cysteine proteinases and their protein inhibitors. Biol. Chem. 378: 141–150 27 Guncar G., Pungercic G., Klemencic I., Turk V. and Turk D. (1999) Crystal structure of MHC class II-associated p41 Ii fragment bound to cathepsin L reveals the structural basis for differentiation between cathepsins L and S. EMBO J. 18: 793–803 28 Rigden D. J., Mosolov V. V. and Galperin M. Y. (2002) Sequence conservation in the chagasin family suggests a common trend in cysteine proteinase binding by unrelated protein inhibitors. Protein Sci. 11: 1971–1977 29 Chai J., Shiozaki E., Srinivasula S. M., Wu Q., Datta P., Alnemri E. S. et al. (2001) Structural basis of caspase-7 inhibition by XIAP. Cell 104: 769–780 30 Huang Y., Park Y. C., Rich R. L., Segal D., Myszka D. G. and Wu H. (2001) Structural basis of caspase inhibition by XIAP: differential roles of the linker versus the BIR domain. Cell 104: 781–790 31 Riedl S. J., Renatus M., Schwarzenbacher R., Zhou Q., Sun C., Fesik S. W. et al. (2001) Structural basis for the inhibition of caspase-3 by XIAP. Cell 104: 791–800 32 Xu G., Cirilli M., Huang Y., Rich R. L., Myszka D. G. and Wu H. (2001) Covalent inhibition revealed by the crystal structure of the caspase- 8/p35 complex. Nature 410: 494– 497 Protein inhibitors of serine proteases 33 Zhou Q. and Salvesen G. S. (2000) Viral caspase inhibitors CrmA and p35. Methods Enzymol. 322: 143 – 154 34 Simonovic M., Gettins P. G. W. and Volz K. (2000) Crystal structure of viral serpin crmA provides insights into its mechanism of cysteine proteinase inhibition. Protein Sci. 9: 1423 – 1427 35 Rees D. C. and Lipscomb W. N. (1982) Refined crystal structure of the potato inhibitor complex of carboxypeptidase A at 2.5 Å resolution. J. Mol. Biol. 160: 475 – 498 36 Seeram S. S., Hiraga K. and Oda K. (1997) Resynthesis of reactive site peptide bond and temporary inhibition of Streptomyces metalloproteinase inhibitor. J. Biochem. (Tokyo) 122: 788 – 794 37 Fernandez-Catalan C., Bode W., Huber R., Turk D., Calvete J. J., Lichte A. et al. (1998) Crystal structure of the complex formed by the membrane type 1-matrix metalloproteinase with the tissue inhibitor of metalloproteinases-2, the soluble progelatinase A receptor. EMBO J. 17: 5238 – 5248 38 Gomis-Ruth F. X., Maskos K., Betz M., Bergner A., Huber R., Suzuki K. et al. (1997) Mechanism of inhibition of the human matrix metalloproteinase stromelysin-1 by TIMP-1. Nature 389: 77 – 81 39 Hege T., Feltzer R. E., Gray R. D. and Baumann U. (2001) Crystal structure of a complex between Pseudomonas aeruginosa alkaline protease and its cognate inhibitor: inhibition by a zincNH2 coordinative bond. J. Biol. Chem. 276: 35087– 35092 40 Li M., Phylip L. H., Lees W. E., Winther J. R., Dunn B. M., Wlodawer A. et al. (2000) The aspartic proteinase from Saccharomyces cerevisiae folds its own inhibitor into a helix. Nat. Struct. Biol. 7: 113 – 117 41 Ng K. K., Petersen J. F., Cherney M. M., Garen C., Zalatoris J. J., Rao-Naik C. et al. (2000) Structural basis for the inhibition of porcine pepsin by Ascaris pepsin inhibitor-3. Nat. Struct. Biol. 7: 653 – 657 42 Janin J. and Chothia C. (1990) The structure of protein-protein recognition sites. J. Biol. Chem. 265: 16027 – 16030 43 Apostoluk W. and Otlewski J. (1998) Variability of the canonical loop conformations in serine proteinases inhibitors and other proteins. Proteins 32: 459 – 474 44 Schechter I. and Berger A. (1967) On the size of the active site in proteases. I. Papain. Biochem. Biophys. Res. Commun. 27: 157 – 162 45 Jackson R. M. and Russell R. B. (2000) The serine protease inhibitor canonical loop conformation: examples found in extracellular hydrolases, toxins, cytokines and viral proteins. J. Mol. Biol. 296: 325 – 334 46 Korsinczky M. L., Schirra H. J., Rosengren K. J., West J., Condie B. A., Otvos L. et al. (2001) Solution structures by 1H NMR of the novel cyclic trypsin inhibitor SFTI-1 from sunflower seeds and an acyclic permutant. J. Mol. Biol. 311: 579 – 591 47 Felizmenio-Quimio M. E., Daly N. L. and Craik D. J. (2001) Circular proteins in plants: solution structure of a novel macrocyclic trypsin inhibitor from Momordica cochinchinensis. J. Biol. Chem. 276: 22875 – 22882 48 Goldenberg D. P. and Creighton T. E. (1983) Circular and circularly permuted forms of bovine pancreatic trypsin inhibitor. J. Mol. Biol. 165: 407 – 413 49 Botos I., Wu Z., Lu W. and Wlodawer A. (2001) Crystal structure of a cyclic form of bovine pancreatic trypsin inhibitor. FEBS Lett. 509: 90 – 94 50 Lu W., Apostol I., Qasim M. A., Warne N., Wynn R., Zhang W. L. et al. (1997) Binding of amino acid side-chains to S1 cavities of serine proteinases. J. Mol. Biol. 266: 441 – 461 51 Hohenester E., Maurer P. and Timpl R. (1997) Crystal structure of a pair of follistatin-like and EF-hand calcium-binding domains in BM-40. EMBO J. 16: 3778 – 3786 52 Bode W., Greyling H. J., Huber R., Otlewski J. and Wilusz T. (1989) The refined 2.0 Å X-ray crystal structure of the com- CMLS, Cell. Mol. Life Sci. 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 Vol. 60, 2003 plex formed between bovine beta-trypsin and CMTI-I, a trypsin inhibitor from squash seeds (Cucurbita maxima): topological similarity of the squash seed inhibitors with the carboxypeptidase A inhibitor from potatoes. FEBS Lett. 242: 285–292 Pritchard L. and Dufton M. J. (1999) Evolutionary trace analysis of the Kunitz/BPTI family of proteins: functional divergence may have been based on conformational adjustment. J. Mol. Biol. 285: 1589–1607 Zweckstetter M., Czisch M., Mayer U., Chu M. L., Zinth W., Timpl R. et al. (1996) Structure and multiple conformations of the kunitz-type domain from human type VI collagen alpha3(VI) chain in solution. Structure 4: 195–209 Kohfeldt E., Gohring W., Mayer U., Zweckstetter M., Holak T. A., Chu M. L. et al. (1996) Conversion of the Kunitztype module of collagen VI into a highly active trypsin inhibitor by site-directed mutagenesis. Eur. J. Biochem. 238: 333 – 340 Locht A. van de, Stubbs M. T., Bode W., Friedrich T., Bollschweiler C., Hoffken W. et al. (1996) The ornithodorinthrombin crystal structure, a key to the TAP enigma? EMBO J. 15: 6011–6017 Waxman L., Smith D. E., Arcuri K. E. and Vlasuk G. P. (1990) Tick anticoagulant peptide (TAP) is a novel inhibitor of blood coagulation factor Xa. Science 248: 593–596 Wei A., Alexander R. S., Duke J., Ross H., Rosenfeld S. A. and Chang C. H. (1998) Unexpected binding mode of tick anticoagulant peptide complexed to bovine factor Xa. J. Mol. Biol. 283: 147–154 Moses E. and Hinz H. J. (1983) Basic pancreatic trypsin inhibitor has unusual thermodynamic stability parameters. J. Mol. Biol. 170: 765–776 Makhatadze G. I., Kim K. S., Woodward C. and Privalov P. L. (1993) Thermodynamics of BPTI folding. Protein Sci. 2: 2028–2036 Wieczorek M., Otlewski J., Cook J., Parks K., Leluk J., Wilimowska-Pelc A. et al. (1985) The squash family of serine proteinase inhibitors: amino acid sequences and association equilibrium constants of inhibitors from squash, summer squash, zucchini, and cucumber seeds. Biochem. Biophys. Res. Commun. 126: 646–652 Otlewski J. and Laskowski M. Jr (1985) Calorimetric investigation of the reactive site of turkey ovomucoid third domain. Fed. Proc. Fed. Am. Soc. Exp. Biol. 44: 1807 Krokoszynska I. and Otlewski J. (1996) Thermodynamic stability effects of single peptide bond hydrolysis in protein inhibitors of serine proteinases. J. Mol. Biol. 256: 793– 802 Socorro M. C. M. do, Oliva M. L., Fritz H., Jochum M., Mentele R., Sampaio M. et al. (2002) Characterization of a Kunitz trypsin inhibitor with one disulfide bridge purified from Swartzia pickellii. Biochem. Biophys. Res. Commun. 291: 635–639 Hemmi H., Kumazaki T., Yamazaki T., Kojima S., Yoshida T., Kyogoku Y. et al. (2003) Inhibitory specificity change of the ovomucoid third domain of the silver pheasant upon introduction of an engineered Cys14-Cys39 bond. Biochemistry 42: 2524–2534 Hurle M. R., Marks C. B., Kosen P. A., Anderson S. and Kuntz I. D. (1990) Denaturant-dependent folding of bovine pancreatic trypsin inhibitor mutants with two intact disulfide bonds. Biochemistry 29: 4410–4419 Krokoszynska I., Dadlez M. and Otlewski J. (1998) Structure of single-disulfide variants of bovine pancreatic trypsin inhibitor (BPTI) as probed by their binding to bovine betatrypsin. J. Mol. Biol. 275: 503–513 Rolka K., Kupryszewski G., Rozycki J., Ragnarsson U., Zbyryt T. and Otlewski J. (1992) New analogues of Cucurbita maxima trypsin inhibitor III (CMTI III) with simplified structure. Biol. Chem. Hoppe Seyler 373: 1055–1060 Review Article 2441 69 Yu M. H., Weissman J. S. and Kim P. S. (1995) Contribution of individual side-chains to the stability of BPTI examined by alanine-scanning mutagenesis. J. Mol. Biol. 249: 388 – 397 70 Tamura A., Kanaori K., Kojima S., Kumagai I., Miura K. and Akasaka K. (1991) Mechanisms of temporary inhibition in Streptomyces subtilisin inhibitor induced by an amino acid substitution, tryptophan 86 replaced by histidine. Biochemistry 30: 5275 – 5286 71 Beeser S. A., Goldenberg D. P. and Oas T. G. (1997) Enhanced protein flexibility caused by a destabilizing amino acid replacement in BPTI. J. Mol. Biol. 269: 154 – 164 72 Xie Z. W., Luo M. J., Xu W. F. and Chi C. W. (1997) Two reactive site locations and structure-function study of the arrowhead proteinase inhibitors, A and B, using mutagenesis. Biochemistry 36: 5846 – 5852 73 Ardelt W. and Laskowski M. Jr (1985) Turkey ovomucoid third domain inhibits eight different serine proteinases of varied specificity on the same ...Leu18-Glu19... reactive site. Biochemistry 24: 5313 – 5320 74 Petersen L. C., Bjorn S. E., Olsen O. H., Nordfang O., Norris F. and Norris K. (1996) Inhibitory properties of separate recombinant Kunitz-type-protease-inhibitor domains from tissue-factor-pathway inhibitor. Eur. J. Biochem. 235: 310 – 316 75 Xu Y., Carr P. D., Guss J. M. and Ollis D. L. (1998) The crystal structure of bikunin from the inter-alpha-inhibitor complex: a serine protease inhibitor with two Kunitz domains. J. Mol. Biol. 276: 955 – 966 76 Schirra H. J., Scanlon M. J., Lee M. C., Anderson M. A. and Craik D. J. (2001) The solution structure of C1-T1, a two-domain proteinase inhibitor derived from a circular precursor protein from Nicotiana alata. J. Mol. Biol. 306: 69 – 79 77 Locht A. van de, Lamba D., Bauer M., Huber R., Friedrich T., Kroger B. et al. (1995) Two heads are better than one: crystal structure of the insect derived double domain Kazal inhibitor rhodniin in complex with thrombin. EMBO J. 14: 5149 – 5157 78 McGrath M. E., Erpel T., Bystroff C. and Fletterick R. J. (1994) Macromolecular chelation as an improved mechanism of protease inhibition: structure of the ecotin-trypsin complex. EMBO J. 13: 1502 – 1507 79 Eggers C. T., Wang S. X., Fletterick R. J. and Craik C. S. (2001) The role of ecotin dimerization in protease inhibition. J. Mol. Biol. 308: 975 – 991 80 Trexler M., Banyai L. and Patthy L. (2001) A human protein containing multiple types of protease-inhibitory modules. Proc. Natl. Acad. Sci. USA 98: 3705 – 3709 81 Castro M. J. and Anderson S. (1996) Alanine point-mutations in the reactive region of bovine pancreatic trypsin inhibitor: effects on the kinetics and thermodynamics of binding to betatrypsin and alpha-chymotrypsin. Biochemistry 35: 11435 – 11446 82 Buczek O., Koscielska-Kasprzak K., Krowarsch D., Dadlez M. and Otlewski J. (2002) Analysis of serine proteinase-inhibitor interaction by alanine shaving. Protein Sci. 11: 806 – 819 83 Laskowski M. and Qasim M. A. (2000) What can the structures of enzyme-inhibitor complexes tell us about the structures of enzyme substrate complexes? Biochim. Biophys. Acta 1477: 324 – 337 84 Ardelt W. and Laskowski M. Jr (1991) Effect of single amino acid replacements on the thermodynamics of the reactive site peptide bond hydrolysis in ovomucoid third domain. J. Mol. Biol. 220: 1041 – 1053 85 Siekmann J., Wenzel H. R., Matuszak E., Goldammer E. von and Tschesche H. (1988) The pH dependence of the equilibrium constant KHyd for the hydrolysis of the Lys15-Ala16 reactive-site peptide bond in bovine pancreatic trypsin inhibitor (aprotinin). J. Protein Chem. 7: 633 – 640 86 Otlewski J. and Zbyryt T. (1994) Single peptide bond hydrolysis/resynthesis in squash inhibitors of serine proteinases. 1. 2442 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 D. Krowarsch et al. Kinetics and thermodynamics of the interaction between squash inhibitors and bovine beta-trypsin. Biochemistry 33: 200–207 Musil D., Bode W., Huber R., Laskowski M. Jr, Lin T. Y. and Ardelt W. (1991) Refined X-ray crystal structures of the reactive site modified ovomucoid inhibitor third domains from silver pheasant (OMSVP3*) and from Japanese quail (OMJPQ3*). J. Mol. Biol. 220: 739–755 Betzel C., Dauter Z., Genov N., Lamzin V., Navaza J., Schnebli H. P. et al. (1993) Structure of the proteinase inhibitor eglin c with hydrolysed reactive centre at 2.0 Å resolution. FEBS Lett. 317: 185–188 Shaw G. L., Davis B., Keeler J. and Fersht A. R. (1995) Backbone dynamics of chymotrypsin inhibitor 2: effect of breaking the active site bond and its implications for the mechanism of inhibition of serine proteases. Biochemistry 34: 2225 – 2233 Liu J., Prakash O., Huang Y., Wen L., Wen J. J., Huang J. K. et al. (1996) Internal mobility of reactive-site-hydrolyzed recombinant Cucurbita maxima trypsin inhibitor-V characterized by NMR spectroscopy: evidence for differential stabilization of newly formed C- and N-termini. Biochemistry 35: 12503–12510 Krishnamoorthi R., Lin C. L. and VanderVelde D. (1992) Structural consequences of the natural substitution, E9K, on reactive-site-hydrolyzed squash (Cucurbita maxima) trypsin inhibitor (CMTI), as studied by two-dimensional NMR. Biochemistry 31: 4965–4969 Laskowski M. Jr and Sealock R. W. (1971) Protein proteinase inhibitors – molecular aspects. In: the Enzymes, vol. 3, pp. 376 – 457, Boyer P. D. (ed.), Academic Press, New York Cai M., Gong Y., Prakash O. and Krishnamoorthi R. (1995) Reactive-site hydrolyzed Cucurbita maxima trypsin inhibitorV: function, thermodynamic stability, and NMR solution structure. Biochemistry 34: 12087–12094 Laskowski M. Jr, Kato I., Ardelt W., Cook J., Denton A., Empie M. W. et al. (1987) Ovomucoid third domains from 100 avian species: isolation, sequences, and hypervariability of enzyme-inhibitor contact residues. Biochemistry 26: 202 – 221 Otlewski J. and Krowarsch D. (1996) Squash inhibitor family of serine proteinases. Acta Biochim. Pol. 43: 431 – 444 Beuning L. L., Spriggs T. W. and Christeller J. T. (1994) Evolution of the proteinase inhibitor I family and apparent lack of hypervariability in the proteinase contact loop. J. Mol. Evol. 39: 644–654 Nielsen K. J., Alewood D., Andrews J., Kent S. B. and Craik D. J. (1994) An 1H NMR determination of the three-dimensional structures of mirror-image forms of a Leu-5 variant of the trypsin inhibitor from Ecballium elaterium (EETI-II). Protein Sci. 3: 291–302 Heinz D. W., Hyberts S. G., Peng J. W., Priestle J. P., Wagner G. and Grutter M. G. (1992) Changing the inhibitory specificity and function of the proteinase inhibitor eglin c by sitedirected mutagenesis: functional and structural investigation. Biochemistry 31: 8755–8766 Grzesiak A., Helland R., Smalas A. O., Krowarsch D., Dadlez M. and Otlewski J. (2000) Substitutions at the P(1) position in BPTI strongly affect the association energy with serine proteinases. J. Mol. Biol. 301: 205–217 Brauer A. B., Domingo G. J., Cooke R. M., Matthews S. J. and Leatherbarrow R. J. (2002) A conserved cis peptide bond is necessary for the activity of Bowman-Birk inhibitor protein. Biochemistry 41: 10608–10615 Cierpicki T. and Otlewski J. (2000) Determination of a high precision structure of a novel protein, Linum usitatissimum trypsin inhibitor (LUTI), using computer-aided assignment of NOESY cross-peaks. J. Mol. Biol. 302: 1179–1192 Cai M., Huang Y., Prakash O., Wen L., Dunkelbarger S. P., Huang J. K. et al. (1996) Differential modulation of binding loop flexibility and stability by Arg50 and Arg52 in Cucurbita Protein inhibitors of serine proteases 103 104 105 106 107 108 109 110 111 112 113 114 115 116 maxima trypsin inhibitor-V deduced by trypsin-catalyzed hydrolysis and NMR spectroscopy. Biochemistry 35: 4784 – 4794 Cai M., Gong Y. X., Wen L. and Krishnamoorthi R. (2002) Correlation of binding-loop internal dynamics with stability and function in potato I inhibitor family: relative contributions of Arg(50) and Arg(52) in Cucurbita maxima trypsin inhibitor-V as studied by site-directed mutagenesis and NMR spectroscopy. Biochemistry 41: 9572 – 9579 Fujinaga M., Sielecki A. R., Read R. J., Ardelt W., Laskowski M. Jr and James M. N. (1987) Crystal and molecular structures of the complex of alpha-chymotrypsin with its inhibitor turkey ovomucoid third domain at 1.8 Å resolution. J. Mol. Biol. 195: 397 – 418 Grasberger B. L., Clore G. M. and Gronenborn A. M. (1994) High-resolution structure of Ascaris trypsin inhibitor in solution: direct evidence for a pH-induced conformational transition in the reactive site. Structure 2: 669 – 678 Cierpicki T., Bania J. and Otlewski J. (2000) NMR solution structure of Apis mellifera chymotrypsin/cathepsin G inhibitor-1 (AMCI-1): structural similarity with Ascaris protease inhibitors. Protein Sci. 9: 976 – 984 Osmark P., Sorensen P. and Poulsen F. M. (1993) Context dependence of protein secondary structure formation: the threedimensional structure and stability of a hybrid between chymotrypsin inhibitor 2 and helix E from subtilisin Carlsberg. Biochemistry 32: 11007 – 11014 Cierpicki T. and Otlewski J. (2002) NMR structures of two variants of bovine pancreatic trypsin inhibitor (BPTI) reveal unexpected influence of mutations on protein structure and stability. J. Mol. Biol. 321: 647 – 658 Seeram S. S., Hiraga K., Saji A., Tashiro M. and Oda K. (1997) Identification of reactive site of a proteinaceous metalloproteinase inhibitor from Streptomyces nigrescens TK-23. J. Biochem. (Tokyo) 121: 1088 – 1095 Finkenstadt W. R. and Laskowski M. Jr (1967) Resynthesis by trypsin of the cleaved peptide bond in modified soybean trypsin inhibitor. J. Biol. Chem. 242: 771 – 773 Estell D. A., Wilson K. A. and Laskowski M. Jr (1980) Thermodynamics and kinetics of the hydrolysis of the reactive-site peptide bond in pancreatic trypsin inhibitor (Kunitz) by Dermasterias imbricata trypsin 1. Biochemistry 19: 131 – 137 Ardelt W. and Laskowski M. Jr (1983) Thermodynamics and kinetics of the hydrolysis and resynthesis of the reactive site peptide bond in turkey ovomucoid third domain by aspergillopeptidase B. Acta Biochim. Pol. 30: 115 – 126 Huber R., Kukla D., Bode W., Schwager P., Bartels K., Deisenhofer J. et al. (1974) Structure of the complex formed by bovine trypsin and bovine pancreatic trypsin inhibitor. II. Crystallographic refinement at 1.9 Å resolution. J. Mol. Biol. 89: 73 – 101 Scheidig A. J., Hynes T. R., Pelletier L. A., Wells J. A. and Kossiakoff A. A. (1997) Crystal structures of bovine chymotrypsin and trypsin complexed to the inhibitor domain of Alzheimer’s amyloid beta-protein precursor (APPI) and basic pancreatic trypsin inhibitor (BPTI): engineering of inhibitors with altered specificities. Protein Sci. 6: 1806 – 1824 Chen Z. and Bode W. (1983) Refined 2.5 Å X-ray crystal structure of the complex formed by porcine kallikrein A and the bovine pancreatic trypsin inhibito: crystallization, Patterson search, structure determination, refinement, structure and comparison with its components and with the bovine trypsinpancreatic trypsin inhibitor complex. J. Mol. Biol. 164: 283 – 311 Locht A. van de, Bode W., Huber R., Le Bonniec B. F., Stone S. R., Esmon C. T. et al. (1997) The thrombin E192Q-BPTI complex reveals gross structural rearrangements: implications for the interaction with antithrombin and thrombomodulin. EMBO J. 16: 2977 – 2984 CMLS, Cell. Mol. Life Sci. Vol. 60, 2003 117 Zhang E., St Charles R. and Tulinsky A. (1999) Structure of extracellular tissue factor complexed with factor VIIa inhibited with a BPTI mutant. J. Mol. Biol. 285: 2089–2104 118 Bode W., Schwager P. and Huber R. (1978) The transition of bovine trypsinogen to a trypsin-like state upon strong ligand binding: the refined crystal structures of the bovine trypsinogenpancreatic trypsin inhibitor complex and of its ternary complex with Ile-Val at 1.9 Å resolution. J. Mol. Biol. 118: 99–112 119 Helland R., Otlewski J., Sundheim O., Dadlez M. and Smalas A. O. (1999) The crystal structures of the complexes between bovine beta-trypsin and ten P1 variants of BPTI. J. Mol. Biol. 287: 923–942 120 Finkenstadt W. R., Hamid M. A., Mattis J. A., Schrode J. A., Sealock R. W. and Laskowski M. J. (1974) Kinetics and thermodynamics of the interaction of proteinases with protein inhibitors. In: Bayer-Symposium V, pp. 389 – 411, Fritz H., Tschesche H., Greene L. J. and Truscheit E. (eds), Springer, Berlin 121 Buczek O., Krowarsch D. and Otlewski J. (2002) Thermodynamics of single peptide bond cleavage in bovine pancreatic trypsin inhibitor (BPTI). Protein Sci. 11: 924–932 122 Radisky E. S. and Koshland D. E. Jr (2002) A clogged gutter mechanism for protease inhibitors. Proc. Natl. Acad. Sci. USA 99: 10316–10321 123 Lu W., Qasim M. A., Laskowski M. Jr and Kent S. B. (1997) Probing intermolecular main chain hydrogen bonding in serine proteinase-protein inhibitor complexes: chemical synthesis of backbone-engineered turkey ovomucoid third domain. Biochemistry 36: 673–679 124 McPhalen C. A. and James M. N. (1988) Structural comparison of two serine proteinase-protein inhibitor complexes: eglin-c-subtilisin Carlsberg and CI-2-subtilisin Novo. Biochemistry 27: 6582–6598 125 Takeuchi Y., Satow Y., Nakamura K. T. and Mitsui Y. (1991) Refined crystal structure of the complex of subtilisin BPN¢ and Streptomyces subtilisin inhibitor at 1.8 Å resolution. J. Mol. Biol. 221: 309–325 126 Lin G., Bode W., Huber R., Chi C. and Engh R. A. (1993) The 0.25-nm X-ray structure of the Bowman-Birk-type inhibitor from mung bean in ternary complex with porcine trypsin. Eur. J. Biochem. 212: 549–555 127 Bateman K. S., Huang K., Anderson S., Lu W., Qasim M. A., Laskowski M. Jr et al. (2001) Contribution of peptide bonds to inhibitor-protease binding: crystal structures of the turkey ovomucoid third domain backbone variants OMTKY3-Pro18I and OMTKY3-psi[COO]-Leu18I in complex with Streptomyces griseus proteinase B (SGPB) and the structure of the free inhibitor, OMTKY-3-psi[CH2NH2+]-Asp19I. J. Mol. Biol. 305: 839–849 128 Peng J. W. and Wagner G. (1992) Mapping of the spectral densities of N-H bond motions in eglin c using heteronuclear relaxation experiments. Biochemistry 31: 8571–8586 129 Bode W. and Huber R. (1978) Crystal structure analysis and refinement of two variants of trigonal trypsinogen: trigonal trypsin and PEG (polyethylene glycol) trypsinogen and their comparison with orthorhombic trypsin and trigonal trypsinogen. FEBS Lett. 90: 265–269 130 Bode W. (1979) The transition of bovine trypsinogen to a trypsin-like state upon strong ligand binding. II. The binding of the pancreatic trypsin inhibitor and of isoleucine-valine and of sequentially related peptides to trypsinogen and to p-guanidinobenzoate-trypsinogen. J. Mol. Biol. 127: 357–374 131 Antonini E., Ascenzi P., Bolognesi M., Gatti G., Guarneri M. and Menegatti E. (1983) Interaction between serine (pro)enzymes, and Kazal and Kunitz inhibitors. J. Mol. Biol. 165: 543–558 132 Zbyryt T. and Otlewski J. (1991) Interaction between squash inhibitors and bovine trypsinogen. Biol. Chem. Hoppe Seyler 372: 255–262 Review Article 2443 133 Laskowski M. Jr (1986) Protein inhibitors of serine proteinases – mechanism and classification. Adv. Exp. Med. Biol. 199: 1 – 17 134 Kojima S., Nishiyama Y., Kumagai I. and Miura K. (1991) Inhibition of subtilisin BPN¢ by reaction site P1 mutants of Streptomyces subtilisin inhibitor. J. Biochem. (Tokyo) 109: 377 – 382 135 Polanowska J., Krokoszynska I., Czapinska H., Watorek W., Dadlez M. and Otlewski J. (1998) Specificity of human cathepsin G. Biochim. Biophys. Acta 1386: 189 – 198 136 Krowarsch D., Dadlez M., Buczek O., Krokoszynska I., Smalas A. O. and Otlewski J. (1999) Interscaffolding additivity: binding of P1 variants of bovine pancreatic trypsin inhibitor to four serine proteases. J. Mol. Biol. 289: 175 – 186 137 Abul Qasim M., Ranjbar M. R., Wynn R., Anderson S. and Laskowski M., Jr. (1995) Ionizable P1 residues in serine proteinase inhibitors undergo large pK shifts on complex formation. J. Biol. Chem. 270: 27419 – 27422 138 Qasim M. A., Ganz P. J., Saunders C. W., Bateman K. S., James M. N. and Laskowski M. Jr (1997) Interscaffolding additivity: association of P1 variants of eglin c and of turkey ovomucoid third domain with serine proteinases. Biochemistry 36: 1598 – 1607 139 Huang K., Lu W., Anderson S., Laskowski M. Jr and James M. N. (1995) Water molecules participate in proteinase-inhibitor interactions: crystal structures of Leu18, Ala18, and Gly18 variants of turkey ovomucoid inhibitor third domain complexed with Streptomyces griseus proteinase B. Protein Sci. 4: 1985 – 1997 140 Helland R., Berglund G. I., Otlewski J., Apostoluk W., Andersen O. A., Willassen N. P. et al. (1999) High-resolution structures of three new trypsin-squash-inhibitor complexes: a detailed comparison with other trypsins and their complexes. Acta Crystallogr. D Biol. Crystallogr. 55: 139 – 148 141 Beckmann J., Mehlich A., Schroder W., Wenzel H. R. and Tschesche H. (1988) Preparation of chemically ‘mutated’ aprotinin homologues by semisynthesis: P1 substitutions change inhibitory specificity. Eur. J. Biochem. 176: 675 – 682 142 Krystek S., Stouch T. and Novotny J. (1993) Affinity and specificity of serine endopeptidase-protein inhibitor interactions: empirical free energy calculations based on X-ray crystallographic structures. J. Mol. Biol. 234: 661 – 679 143 Qasim M. A., Lu S. M., Ding J., Bateman K. S., James M. N., Anderson S. et al. (1999) Thermodynamic criterion for the conformation of P1 residues of substrates and of inhibitors in complexes with serine proteinases. Biochemistry 38: 7142 – 7150 144 Wells J. A. (1990) Additivity of mutational effects in proteins. Biochemistry 29: 8509 – 8517 145 Dill K. A. (1997) Additivity principles in biochemistry. J. Biol. Chem. 272: 701 – 704 146 Lu S. M., Lu W., Qasim M. A., Anderson S., Apostol I., Ardelt W. et al. (2001) Predicting the reactivity of proteins from their sequence alone: Kazal family of protein inhibitors of serine proteinases. Proc. Natl. Acad. Sci. USA 98: 1410 – 1415 147 Lu W., Zhang W., Molloy S. S., Thomas G., Ryan K., Chiang Y. et al. (1993) Arg15-Lys17-Arg18 turkey ovomucoid third domain inhibits human furin. J. Biol. Chem. 268: 14583 – 14585 148 Laskowski M. Jr, Qasim M. A. and Yi Z. (2003) Additivitybased prediction of equilibrium constants for some proteinprotein associations. Curr. Opin. Struct. Biol. 13: 130 – 139 149 Stubbs M. T. and Bode W. (1994) Coagulation factors and their inhibitors. Curr. Opin. Struct. Biol. 4: 823 – 832 150 Huntington J. A. and Carrell R. W. (2001) The serpins: nature’s molecular mousetraps. Sci. Prog. 84: 125 – 136 151 Stubbs M. T., Laber B., Bode W., Huber R., Jerala R., Lenarcic B. et al. (1990) The refined 2.4 Å X-ray crystal structure of recombinant human stefin B in complex with the cysteine 2444 152 153 154 155 156 157 D. Krowarsch et al. proteinase papain: a novel type of proteinase inhibitor interaction. EMBO J. 9: 1939–1947 Lenarcic B. and Bevec T. (1998) Thyropins – new structurally related proteinase inhibitors. Biol. Chem. 379: 105 – 111 Stennicke H. R., Ryan C. A. and Salvesen G. S. (2002) Reprieval from execution: the molecular basis of caspase inhibition. Trends. Biochem. Sci. 27: 94–101 Molina M. A., Marino C., Oliva B., Aviles F. X. and Querol E. (1994) C-tail valine is a key residue for stabilization of complex between potato inhibitor and carboxypeptidase A. J. Biol. Chem. 269: 21467–21472 Clore G. M., Gronenborn A. M., Nilges M. and Ryan C. A. (1987) Three-dimensional structure of potato carboxypeptidase inhibitor in solution: a study using nuclear magnetic resonance, distance geometry, and restrained molecular dynamics. Biochemistry 26: 8012–8023 Tate S., Ohno A., Seeram S. S., Hiraga K., Oda K. and Kainosho M. (1998) Elucidation of the mode of interaction of thermolysin with a proteinaceous metalloproteinase inhibitor, SMPI, based on a model complex structure and a structural dynamics analysis. J. Mol. Biol. 282: 435–446 Ohno A., Tate S., Seeram S. S., Hiraga K., Swindells M. B., Oda K. et al. (1998) NMR structure of the Streptomyces met- Protein inhibitors of serine proteases 158 159 160 161 162 alloproteinase inhibitor, SMPI, isolated from Streptomyces nigrescens TK-23: another example of an ancestral beta gammacrystallin precursor structure. J. Mol. Biol. 282: 421 – 433 Baumann U., Bauer M., Letoffe S., Delepelaire P. and Wandersman C. (1995) Crystal structure of a complex between Serratia marcescens metallo-protease and an inhibitor from Erwinia chrysanthemi. J. Mol. Biol. 248: 653 – 661 Gomez D. E., Alonso D. F., Yoshiji H. and Thorgeirsson U. P. (1997) Tissue inhibitors of metalloproteinases: structure, regulation and biological functions. Eur. J. Cell Biol. 74: 111 – 122 Gomis-Ruth F. X., Gomez-Ortiz M., Vendrell J., Ventura S., Bode W., Huber R. et al. (1997) Crystal structure of an oligomer of proteolytic zymogens: detailed conformational analysis of the bovine ternary complex and implications for their activation. J. Mol. Biol. 269: 861 – 880 Dreyer T., Valler M. J., Kay J., Charlton P. and Dunn B. M. (1985) The selectivity of action of the aspartic-proteinase inhibitor IA3 from yeast (Saccharomyces cerevisiae). Biochem. J. 231: 777 – 779 Kageyama T. (1998) Molecular cloning, expression and characterization of an Ascaris inhibitor for pepsin and cathepsin E. Eur. J. Biochem. 253: 804 – 809