Proteinase inhibitors from the european medicinal leech Hirudo medicinalis: Structural, functional and biomedical aspects

Pergamon Molec. Aspects Med. Vol. 16, pp. 215-313, 1995

Copyright 0 1995 Elsevier Science Ltd Printed in Great Britain. All riahts reserved.

0098--2997(95)00002-x 0098-2is97195 $29.00

Proteinase Inhibitors from the European Medicinal Leech Himdo medicinalis Structural, Functional and

Biomedical Aspects*

Paolo Ascenzit$, Gino Amiconi§, Wolfram Boden, Martin0 Bolognesi,l Massimo Coletta** and Enea Menegatti***

#Department of Pharmaceutical Chemistry and Technology, University of Torino, Via Pie tro Giuria 9, 10125 Torino, Italy

$CNR, Center for Molecular Biology, and Department of Biochemical Sciences ‘Alessandro Rossi Fanelli’, University of Rome ‘La Sapienza’, Piazza/e Aldo Moro 5, 00185 Rome, Italy

VMax-Planck-lnstitut fiir Biochemie, D-82152 Martinsried bei Miinchen, Gennany [Center for Advanced Biotechnology /ST and Department of Physics, University of Genova,

Viale Benedetto XV 10, 16132 Genova, Italy **Department of Molecular, Cellular and Animal Biology, University of Camerino, Via Filippo

Camerini 2, 62032 Camerino (MC), Italy “‘Department of Pharmaceutical Sciences, University of Ferrara, Via Fossato di Mortara

7 7119, 44 100 Ferrara, Italy

Contents

CHAPTER 1 Introduction 217

CHAPTER 2 Relevance of Research on Proteinase inhibitors 221

CHAPTER 3 Hirudin 231

CHAPTER 4 Hirustasin 249

CHAPTER 5 Eglin 255

This paper is dedicated to Professor Antonio Ascenzi on the occasion of his 80th birthday. tAll correspondence should be sent to Professor Paolo Ascenzi at the following address: CNR, Center

for Molecular Biology, and Department of Biochemical Sciences ‘Alessandro Rossi Fanelli’, University of Rome ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Rome, Italy.

215

216

CHAPTER 6 Bdellin

CHAPTER 7 Ttyptase Inhibitor

CHAPTER 8 Conclusions

ACKNOWLEDGEMENTS

REFERENCES

P. Ascenzi et al.

275

279

283

285

287

Introduction

General Aspects

Many physiological and pathological processes are mediated by proteinases and by the products of their action (such as proteinase-generated peptides). However, the uncontrolled activity of proteinases can be deleterious and dangerous to cells, tissues, organs, and finally, to the whole organism. As a result, a series of highly specific (i.e. regulatory) and non-specific (i.e. protective) proteinase inhibitors are available in biological systems. Sometimes (e.g. during acute inflammation), the equilibrium between blood or tissue proteinases and their inhibitors is altered dramatically, so that a massive release of proteinases from injured cells, infiltrating neutrophils and macrophages occurs locally (Horn and Heidland, 1982). In order to reduce damages, these enzymes are promptly neutralized by a variety of inhibitors present in body fluids, resulting in the formation of adducts, which are presumably eliminated through receptor- mediated endocytosis (Pizzo et al., 1988). In other cases, uncontrolled proteolysis can occur since regulating proteinase inhibitors are reduced in concentration relative to healthy people through an inborn error. The classical example is the development of familial emphysema, in which the serum levels of a,-proteinase inhibitor are reduced significantly due to the inability of the mutant protein to be secreted. In general, there are several ways in vivo to develop an imbalance between proteinase and their inhibitors in favour of uncontrolled proteolytic activity. Thus, in addition to genetic mutations (e.g. such as in the case of a,-proteinase inhibitor), these include inhibitor saturation by massive proteinase release, oxidative inactivation and proteolytic inactivation (Johnson and Travis, 1979; Beatty et al., 1980; Banda et al., 1987).

These arguments are currently at the basis of extensive investigations aimed to the development and/or discovery of proteinase inhibitors useful for application in therapy. One line of research is based on the observation that several animal species actively synthesize salivary anticoagulants, i.e. clotting inhibitors, presumably to secure their nutritional requirements for fluid blood from their prey. Among these animals, such as leeches, ticks, bugs, mosquitoes, snakes, and bats (see Ribeiro and Garcia, 1981; Ribeiro et al., 1985; Dunwiddie et al., 1993; Garde11 and Friedman, 1993), leeches have been investigated extensively for their potential utility in preparing, or

217

218 P. Ascenzi et al.

more realistically, in developing drugs of biological origin for therapeutic usage (see Seemtiller et al., 1977; Markwardt, 1988; Stone and Maraganore, 1993).

A Modern Use For An Ancient Remedy

Leeches, annelids belonging to the class Hirudinea, have been utilized in Western medicine through the centuries to treat a multitude of physical complaints. The first writer to mention the medicinal use of the leech was Nicander of Colophain (185-135 Bc), a Greek physician and poet (Nicander of Colophain, 1499). A few centuries later, Galen (129-199 AC) wrote a treatise on the medicinal leech (De Hirudinibus) as a tool to restore the balance of the four humors (yellow and black bile, phlegm and blood), constituting the human being, altered by disease (Major, 1954). In the nineteenth century, the indications for leeching became legion (Major, 1954). The widespread use of leeches in therapy was due to the prevailing opinion that life depends upon irritation and in particular upon heat, which excites the chemical processes in the body (Major, 1954). Therefore, since nature was credited to have no healing power without stimulation, it was necessary to abort disease by active measures, such as leeching. In France, where most advances in medicine occurred up to the year 1850, Broussais (1772-1838) applied to all his patients as many as 30 to 50 leeches, while five to eight were prescribed in cases of extreme debility (Garrison, 1914). By the end of the nineteenth century, the medical use of leeches did not comply with the emerging modern concepts of medicine and, accordingly, began to wane (Major, 1954).

Although less used than formerly (by the mid 18OOs, the demand for leeches was estimated in tens of millions per year in most of the Western nations), presently, leeches have also staged a medical revival. Thus, microsurgeons use them in skin grafting, for the removal of coagulated blood from beneath the new skin as well as a means of treating irrepairable venous insufficiency in pedicled flaps and free tissue transfers (Wells et al., 1993). In particular, these parasites have proved invaluable in plastic and reconstructive surgery, where arterial input can be established, and yet, for some technical reason, veins are not available (Foucher et al., 1981; Batchelor et al., 1984; Lim, 1986). Moreover, leeches are often the least painful way to reduce inflammation (Wells et al., 1993).

Some Anatomical and Physiological Aspects of Leech

Many leeches are the blood-sucking species within the class Hirudinea, e.g. Hirudo medicinalti, Haementeria ghilianii, Haementeria lutzi, Haementeria ojj%5n&, Hirudinaria granulosa, Hirudinaria manillensis, Herpobdella punctata and Nephelipsis obscura. The classical European medicinal leech is, however, Hirudo medicinaiis, a 610 cm long annelid that produces the deepest bite with the longest period of blood extravasion. Placed in contact with skin, this leech bites with the small sucker at the anterior end where three radiating jaws are and draws approximately 5-12 ml of blood. However, the therapeutic effect (the reader is reminded that the primary indication for leeching is venous insufficiency) is not due to the volume of blood ingested by the leech, but rather to the continued bleeding after the removing of the parasite from the host. This phenomenon is due to anticoagulant properties of the leech salivary gland secretion (Engemann and Hegner, 1981; Adams, 1988; Bates et al., 1989; Wells et al., 1993).

Proteinase Inhibitors from Leech Hhdo medicinalis 219

A Brief Account of Modern Biomedical Research on Leech

More then 100 years ago (in 1884) John B. Haycraft, Professor of Physiology in Birmingham, who was then working in Schmiedeberg’s pharmacological laboratory in Strasbourg, discovered the presence of an anticoagulant substance in the leech, demonstrating that the active compound was: (1) present only in the leech head; (2) soluble in water, but not in ethanol or chloroform; (3) to some extent specific, since it prevented blood clotting, but not the coagulation of milk; (4) relatively non-toxic, although rabbits and dogs treated with aqueous leech extracts were a bit proven, shortly after the intravenous injection; and (5) excreted by the kidney (Markwardt and Landmann, 1971; Starke, 1989). In 1905, the active principle present in stable dry extracts from Hirudo medicinalis was named hirudin (Bodong, 1905). Up to the discovery of heparin, hirudin was the only means to prevent blood from clotting. Attempts to isolate the anticoagulant agent and to characterize its site of action were only successful after protein chemistry had developed and the biochemistry of blood coagulation had been elucidated. Markwardt (1957) was the first to prepare pure hirudin and to analyze its mechanism of action demonstrating that it was a thrombin inhibitor. The antithrombotic effect of hirudin was demonstrated in several thrombosis models (Markwardt, 1959; Wallis, 1988), but its clinical use remained limited because of the low amounts available and of the recent scarcity of medicinal leeches. Beyond that, these parasites have been placed recently on the list of endangered species (Markwardt, 1988). However, expression of a recombinant gene for hirudin has been achieved recently by several groups in both yeast and bacteria with high yield (e.g. Harvey et al., 1986; Marki et al., 1991). Therefore, the large scale production of hirudin (and other leech-derived proteins) by genetic engineering is a reality.

Later, from the 1970s onwards, it was found that leech extracts exhibit many other inhibitory activities against different serine proteinases. The purified principles were named in various ways: bdellin (mainly acting on plasmin, trypsin and sperm acrosin; Fritz et al., 1969), eglin (strongly inhibiting elastase; Seemtiller et al., 1977); hirustasin (selectively binding to tissue kallikreins; Sdllner et al., 1994); and the leech-derived tryptase inhibitor (effectively blocking the tryptase activity; Sommerhoff et al., 1994). All these inhibitors, together with enzymes such as hyaluronidase, collagenase and ATP-diphosphohydrolase (Rigbi et al., 1987a), present in leech saliva make such composite fluid a system able to maintain the blood in the liquid state during ingestion and storage, to prevent inflammation and to suppress the host’s immune response. This is a general concept emerged from comparative studies on proteins from salivary glands or saliva of bats, insects, snakes as well as leeches (Ribeiro and Garcia, 1981; Ribeiro et al., 1985).

Although this review concentrates on the most recent (after 1985) developments of proteinase inhibitors from the European medicinal leech Hirudo medicinafis, sufficient background information has been included so as to place this new advancement in knowledge into perspective for the unfamiliar reader. Extensive reviews by Schnebli and Braun (1986) and Seemtiller et al. (1986) may be consulted for a detailed discussion of the literature on this topic before 1985.

Chapter 2

Relevance of Research on Proteinase Inhibitors

General Aspects

Generally, macromolecular proteinase inhibitors act competitively with natural substrates. In fact, inhibition occurs as a consequence of binding of the inhibitor reactive site (i.e. of the substrate-like region present on the surface of the inhibitor) to the proteinase active centre (i.e. the substrate binding region present on the enzyme surface) (Cheronis and Repine, 1993). The inhibitor binding to the proteinase active centre may occur in a substrate or product-like manner. In the case of the substrate-like association, the intra- and intermolecular interactions of the inhibitor reactive site with both the inhibitor core (through spacer elements) and the enzyme binding region, stabilize each other and are so tight that decomposition rarely occurs. In the case of product-like binding, the interactions are less strong, but tenacious enough to prevent fast dissociation (Bode and Huber, 1991, 1992). Other serine proteinase inhibitors prevent the access of substrates to the enzyme catalytic centre by binding mainly to surface sites adjacent to the catalytic triad (to the so-called exosites; see Chapter 3): in these cases (hirudin is an excellent example) very high selectivity is achieved (Bode and Huber, 1991, 1992).

The proteinase-inhibitor interaction is essentially second order, and the resulting adduct consists of one molecule of each reactant (with the exception of the so-called multiheaded inhibitors); alternatively, the adduct dissociation may be described by a first-order event (Laskowski and Kato, 1980; Travis and Salvesen, 1983).

In biological systems, a variety of proteinases are in the presence of a variety of inhibitors: if the pathophysiological function of the latter proteins is only protective, a fine recognition of the target enzyme is unnecessary. However, in other cases (e.g. when thrombin inactivation is considered an effective approach to antithrombotic therapy), the enzyme-inhibitor interaction is required to be as specific as possible. Accordingly, among others, there are two relevant aspects to be addressed in the study of proteinase inhibitors: (1) the functional prerequisites for an ideal inhibitor with therapeutic potential (in terms of kinetic and thermodynamic parameters describing the proteinase-inhibitor adduct (de)stabilization); and (2) the ability to form non-covalent complexes of high affinity and specificity, that is the molecular bases of protein-protein recognition.

221


Evaluation of the in viwo Effectiveness of Proteinase Inhibitors

If a proteinase inhibitor is active in vitro, it will not necessarily work in viva. In order to decide whether an inhibitor plays a physiological role or not, three sets of data are required (Bieth, 1974, 1980; Beatty et al., 1980): (1) the in vivo concentration (i.e. the physiological level or the therapeutic concentration), as well as (2) equilibrium quantities and (3) kinetic parameters describing the enzyme-inhibitor adduct formation.

In general, the reaction of a proteinase (P) with a macromolecular inhibitor (r) is not a single-event process, but consists of a spatio-temporal series of interactions. From the phenomenological viewpoint, in fact, the apparent rate constant of the overall process for adduct formation does not increase linearly with the inhibitor concentration, but tends to level off (Quast et al., 1974, 1978; Laskowski and Kato, 1980; Antonini et al., 1983a, b; Ascenzi et al., 1986; Amiconi et al., 1987). Such a behaviour has been interpreted to be indicative of the presence of a relatively fast pre-equilibrium event, associated with the formation of a loose adduct (PI),. This process is followed by a rate-limiting first-order event (due to macromolecular isomerization changes), associated with the transformation of the transient (PI), adduct into the final stable (PI), complex, according to Scheme 1:

Kl k +2

p+z * (P:I), + (P:Z), (1)

k -2

where K, is the association pre-equilibrium constant i.e. the intrinsic affinity constant of the inhibitor for the enzyme, k,, is the rate constant for the isomerization event (representing the rate-limiting pseudo-first-order process), and k_, indicates the rate constant for the reverse of the event described by k+2. The overall association equilibrium constant (K), the overall second-order rate constant for the (P:Z), adduct formation (k,), and the overall enzyme-inhibitor complex dissociation rate constant (k,,,), for the reaction given in Scheme 1, correspond to K = K1.k+2/k_2, k,, = Kl-k+2 and k,, = k_,. The cleavage of the inhibitor at the level of the scissile peptide bond is possible and reversible. In fact, the rate of cleavage is, in most cases, slower than the dissociation rate of the binary adduct which mainly undergoes disaggregation yielding the intact inhibitor and the cleaved form which accumulates only slowly (Quast et al. , 1978).

Under paraphysiological conditions (i.e. at low reactant concentrations) and for all practical purposes, the inhibition process can be described by a simple reaction according to Scheme 2 (Bieth, 1974, 1980; Beatty et al., 1980):

K on

p+z + wo2

k off

(2)

Proteinase Inhibitors from Leech Hirudo medicinalis 223

Even though Scheme 2 is an oversimplification of the reaction pathway of most reversible enzyme inhibitors (see Scheme l), the three macroscopic constants (K, k,, and kOff) are nevertheless valuable parameters for judging the possible physiological role of inhibitors (Bieth, 1974, 1980; Beatty et al., 1980).

For high efficiency in vivo (i.e. in order to protect against deleterious effects of proteinases), a macromolecular inhibitor must react very quickly with enzymes that are released accidentally (e.g. during diseases) or physiologically; in addition, the association must be irreversible or at least dissociation should occur very slowly; these two properties (fast inhibition and stability of adducts) may simply be described by k,, and kOff, measured in vitro. Thus, under pseudo-first-order conditions (i.e. [Z,]>5[P,]), since only the inhibitor in molar excess can play an efficient physiological role, the time required for complete inhibition of a proteinase in vivo (t,) can be estimated (simply recalling textbook physical chemistry (Atkins, 1994)) to be seven times the half-life of the process (i.e. 7.t,,,). Since t,,z = 0.693/(k,;[Z,]),

t, = 7.0.693/(k,, . [Z,,])

Therefore, the t, value may easily be calculated once k,, and the total inhibitor concentration ([Z,]) are known. Even though Eqn 3 is strictly valid only for irreversible inhibitors, in practice, it also holds in vivo for the reversible ones when [Z,].K >>lO* (Bieth, 1974, 1980; Beatty et al., 1980).

In biological systems, the fast formation of the (PI), adduct (see Schemes 1 and 2; related to high k,, values) is desirable. In fact, since the physiological function or the therapeutic activity of a proteinase inhibitor is to protect tissues against baneful injuries, its action is efficient only if the inhibition is quick enough in vivo to prevent the attack to natural and/or potential substrates as much as possible. As a matter of fact, a,-antitrypsin plays an important physiological role in the prevention of emphysema since the value of t, for its interaction with human leukocyte elastase is 3 msec (Beatty et al., 1980). Alternatively, the action of al-antitrypsin against human trypsin (e.g. when massively liberated into blood circulation in the course of acute pancreatitis) is negligibly protective due to the t, value of 21 sec. During this time, in fact, partial activation of both the clotting cascade and of the hypotensive peptide generating proenzymes readily occurs, contributing to induce the so-called pancreatic shock (Beatty et al., 1980).

The competition of a substrate and an inhibitor for a proteinase is negligible when

h%l<<L where [S,] is the total substrate concentration and K,,, is the Michaelis-Menten constant. This effect can easily be taken into account if the system is simple and possibly involves only one substrate molecule. Thus, the apparent association second-order rate constant (k,,‘) is

k 0” * = k,;(K,W + [&J)).

However, when the system is complex (i.e. many proteins interact with the proteinase), the value of k,, should be determined experimentally in order to circumvent this problem (e.g. in the presence of plasma) (Bieth, 1974, 1980; Beatty et al., 1980). It


is also possible to evaluate the extent of substrate hydrolysis during t, for an irreversible inhibitor by applying Eqn 5 (Tian and Tsou, 1982):

F = (k,,&&,,)~[Po]&/7~0.693), (5)

where p is the molar fraction of the substrate hydrolyzed, and kcat is the enzyme catalytic constant for a given substrate. It is noticable that an inhibitor with a rC as high as 15 min may still play a controlling role if [P&1~10-9 M or k,,JK,,,<105 ~-1 set-1 (Bieth, 1974, 1980; Beatty et al., 1980).

High adduct stability (i.e. very low k,, values) is the most desirable property for inhibitors because in these cases a single dose (a high dose in the case of low k,, values) can be sufficient for long periods of time. This can be seen as the preferable choice to ensure efficient inhibition rather than maintaining a high inhibitor concentration over a longer time in the case of high k,, values (Bieth, 1974, 1980; Beatty et al., 1980).

Of course only reversible inhibitors may undergo dissociation from proteinases. If the half-life of the adduct is t,,z = 0.693/k,,, the stability in kinetic terms (tstab, that is the minimal time during which the adduct can be considered stable) may be estimated according to Eqn 6 (Atkins, 1994):

t stab = 0.0693/k,,, (6)

i.e. l/10 of l,,z for the adduct decomposition. The overall stability of the adduct depends also upon the ratio between the inhibitor concentration and K; if [Z,J.K >>103, a reversible inhibitor behaves in uivo like an irreversible one, whatever the proteinase concentration (Bieth, 1974, 1980; Beatty et al., 1980).

The effectiveness of macromolecular proteinase inhibitors has often been evaluated on the basis of this treatment (e.g. Beatty et al., 1980; Onesti et al., 1992).

Recognition Between Protein Surfaces: Swine Proteinase-macromolecular Inhibitor Interaction

The ability to form non-covalent adducts of high affinity and specificity is a basic property of many biological macromolecules. It is through specific interactions that proteins recognize other macromolecules, discriminating them among many others, and spontaneously assemble in stable aggregate(s) and/or structure(s). Such recognition between proteins (that usually implicates a small number, 10 to 30, of amino acid residues on each partner) is at the basis of countless biological processes in which the formation of macromolecular adducts elicits proper key physiological events. Typical examples in this field are the specific interactions between an antigen and an antibody, as well as an enzyme and a macromolecular inhibitor. These events, based on recognition processes, are of complex nature as the formation of a stable adduct relies on several different and concurrent factors, which can be independently considered for the sake of skematization. First, the driving forces that bring the macromolecular partners together, and stabilize them in an adduct which may be very long-lived with respect to the time of typical enzymatic reactions should be considered. According to

Proteinase Inhibitors from Leech Himdo medicinalis 225

the collision theory, and under relatively dilute conditions, rates of association are determined by the amount of time it takes for the diffusing macromolecules initially to encounter each other and the probability that a given encounter will lead to association (Northrup and Erickson, 1992). Related to this is the question of affinity between the different molecular partners which, in order to give rise to a biologically-productive adduct, must also meet criteria related to the specificity of the recognition event. Furthermore, proteins may undergo various levels of conformational transition during the recognition processes, whose energetics are going to influence the overall stability of the protein adduct and the kinetics of its formation (Coletta et al., 1990). As a result, examination of the final binding product evidenced by X-ray crystallography (that is in a time-averaged structure) does not fully define all the reactivity characteristics at the basis of the recognition process exhibited by the serine proteinase-inhibitor system. In fact, the interaction of the two protein surfaces involves molecular flexibility, induced fit, entropy and enthalpy changes, solvent, hydrophobic, van der Waals and electrostatic interactions (Amiconi et al., 1987). Therefore, for a given interacting system, it is highly desirable to have as much information as possible from thermodynamic and kinetic data obtained in various states of the functional events. Some questions can illustrate the kind of information sought: how does each residue at the interface contribute to the energetics of the complex stabilization?; are the pairs of interactions at the contact surface independent of the rest of the molecule, i.e. is the contact like that between two rigid surfaces?; are the interactions modulated by changes in conformation that accompany adduct formation, i.e. does flexibility serve as a regulatory property?; are the interaction dynamics better described with a zipper-like model or with a groping movement?

Molecular surfaces coming into close contact during adduct formation need to posses a given level of structural complementarity both in shape and charge before any recognition process may be started (Bolognesi et al., 1987; Menegatti et al., 1987a). This is the logic consequence of the evidence that no machinery for assembling protein subunits into larger adducts is present in the cells where monomers associate spontaneously (Connolly, 1992). Therefore, complementarity in structural determinants is expected to play a fundamental role in regulating both thermodynamics as well as kinetics of protein-protein complexation. Thus, single residue mutations on inhibitors sharing the same binding loop structure affect stability of different proteinase-inhibitor adducts in different ways (Menegatti ef al., 1987b; Bigler et al., 1993).

A consideration of the entropic cost arising from the loss of degrees of freedom of motion, e.g. lost when two macromolecules are rigidly constrained within an adduct, must be taken into account for a thorough analysis of any protein-protein recognition process. Moreover, if the binding process involves capturing a flexible macromolecule, whose internal rotations about single bonds must also be restricted, the consequence is a further adverse entropic penalty that results in a reduction in the equilibrium dissociation constant (Kd =1/K; see Eqn 5), according to the classical relationships (Atkins, 1994):

6G = - RZlnK, = R’llnK, (7)

and

226 P. Ascenzi eta/.

6G = 6H - T&Y. (8)

The entropy term, calculated in the absence of solvent, represents a destabilizing contribution to the process of adduct formation due to a global entropy decrease; it evaluates the minimum energetic cost which must be paid off in order to reach conditions for the production of the bimolecular complex. Theoretical estimates (Finkelstein and Janin, 1989) evaluate such an entropic contribution to require approximately 15 kcal mol-1 of favourable energetic interactions in order to be compensated. Therefore, in order for productive binding to occur (6G <O), the above adverse cost in bringing about conformational order must be offset by favourable intermolecular interactions, such as hydrogen bonds, van der Waals packing, and in particular, the increased entropy associated with solvent randomization. Relevant for the proteinase-inhibitor complex formation, in fact, is the question of solvent entropy, i.e. consideration of the degrees of freedom lost/gained mainly by water (but also small solute) molecules which are displaced from the interacting protein surfaces, or which come into contact with molecular regions of low polarity. The release of protein surface hydrogen-bonded water molecules to bulk solvent, in general terms, is a process favouring adduct formation through solvent entropy increase; a theoretical evaluation of such a contribution is, however, difficult due to the absence of detailed information on the starting and final states.

The contact areas observed in protein-protein adducts span several hundred A2 of buried protein surface. In all cases, the contact between the interacting surfaces is so tight that the protein-protein contact region is often reminiscent of the atomic packing density found in the protein core regions. Efficient interface packing, which rarely allows the formation of cavities, is achieved through rather contained conformational changes which allow the extensive fulfilment of intermolecular hydrogen-bonded capabilities (Bolognesi et al., 1982, 1987; Frigerio et al., 1992). Comparison of the crystal structures of the free molecular species and of their bimolecular adducts shows that, in general terms, a substantial structural complementarity exists between the interacting surfaces before the recognition event (Bode and Huber, 1991, 1992). Serine proteinase-inhibitor adducts display contact interfaces with an average area of 750 A2 (Janin and Chothia, 1990; Janin, 1995). The largest contact interface observed by far (1800 AZ) is between human a-thrombin and hirudin, the most potent anticlotting agent; it corresponds to approximately 12% of the enzyme available surface (Grtitter et al., 1990; Rydel et al., 1990, 1991; Bode and Huber, 1991, 1992). Half of the interface area derives from the N-terminal domain of hirudin which binds to the enzyme active site much like other proteinase inhibitors do, and the other half from 17 residues present in the hirudin C-terminal tail filling the so-called fibrinogen exosite of thrombin. The C-terminal tail is disordered in the free inhibitor, but it undergoes a disorder-to-order transition upon enzyme binding (Grtitter et al., 1990; Rydel et al., 1990, 1991; Bode and Huber, 1991, 1992). An engineered derivative of hirudin, hirulog, in which much of the N-terminal domain has been deleted, still forms with human or-thrombin a stable adduct with a large interface (approximately 1300 AZ) (Q’ iu et al., 1992). Therefore, the larger than usual interface area in the thrombin-hirudin adduct is needed to compensate for the entropic cost of main chain immobilization. A disorder-to-order transition of comparable relevence is observed when bovine trypsinogen binds to the bovine basic pancreatic trypsin inhibitor (BPTI), as well as to the bovine and porcine pancreatic

Proteinase Inhibitors from Leech Hirudo rr~~#cina/is 227

secretory trypsin inhibitor (PSTI) (Marquart et al., 1983; Amiconi et al., 1987; Coletta

et al., 1990). This phenomenon, however, is reflected not in a change of the interface area, which is the same as in the complex with bovine P-trypsin, but in the affinity whose energy decreases by approximately 8 kcal mol-1 (Marquart et al., 1983; Amiconi et al., 1987; Coletta et al., 1990).

The specificity in recognition between serine proteinases and their protein inhibitors is usually described by particular terminology (Schechter and Berger, 1967). From inspection of the crystal structures of the proteinase-inhibitor adducts, different recognition subsites can be localized at the interacting molecular surfaces. The generally accepted classification (Schechter and Berger, 1967) identifies the proteinase contacting residues as S and S’ according to their location on the N- and C-side of the inhibitor potentially scissile peptide bond, whereas the corresponding specificity recognition subsites present on the inhibitor surface are labelled as P and P’ (see Fig. 1). Additionally, the recognition subsites are numbered sequentially in the two directions, starting from amino-acid residues involved in the potentially scissile peptide bond. Most of the serine proteinase protein inhibitors recognize specific target enzymes through a

Human a-thrombin

S,......Sl Sl . . . . ..s ’

rim

Eglin c 1 -k2 N

Fig. 1. Schematic representation of the human c&hrombixAirudin and bovine achymotrypskxglin c adducts (Bode and Huber, 1991). The terminology of specificity subsites (i.e. S and S’) has been reported only for the two proteinases (Schechter and Berger, 1967). The corresponding binding sites present on the inhibitor surface are represented by teeth (i.e. straight strokes that face the proteinase, and are usually denoted as P and P’ towards the N-terminus and towards the C-terminus, respectively, relative to the enzyme active centre (Schechter and Berger, 1967)). In contrast with the bovine a-chymotrypsin-eglin c adduct (Frigerio et al., 1992), the primary specificity subsite of human a-thrombin is not occupied by hirudin residue(s) (Griitter et al., 1990; Rydel et al., 1990, 1991; Bode and Huber, 1991, 1992). The potentially scissile peptide

bond present in the reactive site of eglin c is shown by a filled square.


non-covalent interaction occurring at the so-called primary specificity subsite, between the P, inhibitor residue and the S, enzyme cleft. In some cases, this interaction is a strong salt link buried in an otherwise low polarity protein environment; in other serine proteinases, this interaction is essentially hydrophobic (see Fig. 1). Alternatively, the S, subsite is unoccupied in the human a-thrombin-hirudin adduct (see Fig. 1) (Grtitter et al., 1990; Rydel et al., 1990, 1991; Huber and Bode, 1991, 1992). Specific recognition occurring at other subsites involves directional interactions, such as hydrogen bonds between main chain as well as side-chain donors or acceptors, which provide juxtaposition of the inhibitor molecule with respect to the serine (pro)enzyme active site residues. As a result of these specific interactions, the contact region of the proteinase-inhibitor adduct is more rigid than in the free species, as indicated by analysis of the crystallographic atomic temperature factors (Bode and Huber, 1991, 1992). Sometimes, interactions between associating proteins occur outside the so-called recognition centre (Schechter and Berger, 1967). Thus, the binding of thrombin to its natural substrate (i.e. fibrinogen) or its potent protein inhibitor (i.e. hirudin) involves the so-called exosite or fibrinogen recognition site (Grtitter et al., 1990; Rydel et ai., 1990, 1991), a domain separate from the active centre of the serine proteinase (see Chapter 3).

Although no detailed mechanism for the recognition between serine (pro)enzymes and their protein inhibitors has been proposed (except for the interaction between thrombin and hirudin or hirulog, as well as bovine l3-trypsin and BPTI and PSTI) (Antonini et al., 1983a, b; Ascenzi et al., 1986; Amiconi et al., 1987; Bolognesi et al., 1987; Ayala and Di Cera, 1994; Parry et al., 1994), the following considerations are in order.

The crystallographic analysis of serine proteinases- macromolecular inhibitor adducts: (1) indicates the existence of multiple points of attachment; (2) proves a rigorous spatial configuration; and (3) attests the presence of many energetically equivalent contributions (Bolognesi et al., 1987). The fact that the binding energy depends upon many small, apparently independent contributions, has some important consequences: (1) since no atomic group in the enzyme or inhibitor is all important, it follows that macromolecular inhibitors differing from one another by a single feature, like the replacement of one or two group(s), will in many cases be bound as well; (2) if protein specificity of binding is the result of many independent effects none of which is dominant, the resulting properties (i.e. the binding energy) will have a normal distribution with respect to those structural changes in the protein that give different weights to the individual interactions; and (3) on this basis, the very origin of molecular recognition implies that it cannot be perfect in any sense (Menegatti et al., 1987a; Janin, 1995).

These conclusions are of general relevance in that catalysis itself appears to be an event delocalized around the active site; in other words, catalysis results not from just a few key residues, but is the consequence of many interactions of small binding energy that conspire to ease the reagents into the transition state and then stabilize the high energy intermediate (Ho and Fersht, 1986; Wells and Fersht, 1986). In addition, flexible segments linked to the recognition centre can affect the binding energy by taking advantage of entropy loss; thus, the system may have high specificity without excessively tight binding (Ringe and Petsko, 1985). Finally, from the kinetic viewpoint, the initial

Proteinase Inhibitors from Leech Hinrdo medicinalis 229

collision between the proteinase and the protein inhibitor has a substantial duration in which the molecules are free to explore each others surfaces (Northrup and Erickson, 1992). However, the movement of the two interacting surfaces is not a smooth steering towards a sequential formation of hydrogen bonds, polar and apolar interactions; during the process, the amino acids relevant to the binding blindly grope along, randomly breaking a salt bridge here, forming a hydrogen bond there, until the system nestles into an energy minimum along the reaction pathway (Amiconi et al., 1987).

Chapter3

Hirudin

General Aspects

Thrombin is a serine proteinase that plays a central role in the coagulation and thrombogenetic processes (Fenton, 1986; Davie et al., 1991; Stubbs and Bode, 1993a, b, 1994, 1995a, b). Thrombin cleaves fibrinogen to form fibrin peptides which then polymerize to form the fibrin clot, subsequently activating Factor XIII to stabilize this polymer network (Janus et al., 1983; Fenton, 1986). The proteinase is also able to activate other blood coagulation factors, such as Factor V, Factor VIII and protein C (Fenton, 1986). In addition, thrombin interacts with cells inducing several types of responses, such as mitosis in fibroblasts and metabolite secretion by platelets (Esmon et al., 1982; Harmon and Jamieson, 1985; Fenton, 1986, 1988).

Different circulating forms of human thrombin have been identified in viva, such as (1) cY-thrombin, which is the native molecule, originating from the N-terminal proteolytic cleavage of prothrombin, consisting of two chains (the A chain with 36 amino acids and the B chain with 259 amino acids) linked through disulfide bonds (Butkowski et al., 1977; Degen et al., 1983); (2) B-thrombin, lacking amino-acid residues between positions 68 and 77A (B-loop) of the B chain; and (3) y-thrombin, from which amino- acid residues between positions 68 and 77A (B-loop), and 127 and 149E (y-loop) of the B chain have been cleaved off (Elion et al., 1986). Other derivatives of the thrombin molecule can be obtained through the proteolytic action of serine proteinases, such as bovine B-trypsin and porcine pancreatic elastase, which induce the formation of &.- and e-thrombin, respectively (Kawabata et al., 1985; Ascenzi et al., 1992a).

Inhibition and modulation of thrombin activity is definitely susceptible of having a widespread action on hemostasis, thus raising a very generalized interest toward the regulation of thrombin action (Berrettini et al., 1987).

The main endogenous inhibitor of thrombin is antithrombin III, which ‘in vivo’ inhibits the proteinase very slowly. Heparin brings about a marked enhancement of the inhibition rate by antithrombin III (Olson and Shore, 1982), this being the reason for its widespread therapeutic utilization as an antithrombotic agent. However, since heparin has a short half-life in the blood and is ineffective in patients with antithrombin

231


III deficiency, from the therapeutical standpoint, the research on alternative exogenous inhibitors of thrombin are sought (Hirsh, 1991).

Among several potential agents, a particularly powerful inhibitor of thrombin activity is hirudin, a small protein extracted and isolated from the salivary glands of the leech Hirudo medicinalis. Hirudin, which has been originally used to inhibit the mitogenic effect of thrombin on fibroblasts (van Obberghen-Schilling et al., 1982) and the activation of platelets (Hoffmann and Markwardt, 1984), has been studied extensively with special reference to its interaction mechanism with thrombin, leading to a detailed understanding of several functional and structural aspects concerning the protein-protein recognition processes underlying its inhibitory action (Haycraft, 1884; Markwardt, 1955, 1986, 1970, 1988, 1992; Seemtiller et al., 1977, 1986; Schnebli and Braun, 1986; Fenton, 1989; Ascenzi et al., 1992a; Stubbs and Bode, 1993a, b, 1994; Banner et al., 1994; Stone, 1995; Verlinde and Hol, 1994).

Hirudin has been isolated not only from the leech Hirudo medicinalis, a species which feeds on frogs and mammals (Seemtiller et al., 1986), but also from Hirudinaria munillensis, a species specialized in feeding on mammals, particularly on water buffaloes (Electricwala et al., 1993a, b; Scacheri et al., 1993). Hirudins belonging to different leech species, Hirudo and Hirudinariu, are closely similar molecules, probably evolved from a common ancestral gene (Electricwala et al., 1993a, b; Scacheri et al., 1993). The hirudin gene isolated from Hirudinariu manillensis contains four exons: the first one corresponds to the signal peptide required for extracellular secretion, while the following three code for the full primary structure of the antithrombin polypeptide (Scacheri et al., 1993).

Structural Aspects

Hirudin may be prepared from the whole leech, from leech heads, and from dilute leech saliva (Bagoly et al., 1976; Seemtiller et al., 1986; Rigbi et al., 1987a). Recombinant hirudin may be obtained by gene expression in Escherichia coli or yeast; at variance with natural hirudin, the Tyr63 residue of the recombinant inhibitor is not sulfated (see Harvey et al., 1986; Marki et al., 1991).

Hirudin exists in multiple forms. Thiee groups of proteins have been identified on the basis of their N-terminal Ile or Thr or Val residues. Valine and isoleucine were found in a ratio of 3:l. Hirudins are single-chain polypeptides with M, values close to 7 kDa, being composed of 65 or 66 amino-acid residues (Dodt et al., 1984; Seemtiller et al., 1986; Mao et al., 1987; Krstenansky et al., 1988a; Tripier, 1988; Scharf et al., 1989). Figure 2 shows the amino-acid sequence of the most abundant hirudin iso-inhibitor variant 1 (hirudin l), and of the iso-inhibitor variant 2, bearing Lys at position 47 (hirudin 2-Lys47), which have been used for NMR solution and X-ray structure investigations (Folkers et al., 1989; Haruyama and Wiithrich, 1989; Griitter et al., 1990; Rydel et al., 1990, 1991). The amino-acid sequence of hirudisin, the recombinant hirudin derivative antagonist of glycoprotein IIb-IIIa as well as inhibitor of platelet aggregation (Knapp et al. 1992; Krezel et al., 1994), is also reported in Fig. 1.

Residues Vall-Pro48 in hirudin 1 (Ilel-Pro48 in hirudin 2-Lys47) are arranged in a looped structure, stabilized by three disulfide bridges (Cys6-Cysl4, Cysl6-Cys28 and

Hirudin

1

Hirudi"

2-Lys47

liirudlsi"

Antistasi"

domain

1

Ghilante"

domain

1

Antistasin

domain

2

Ghilanten

domain

2

Hirustasi"

llec

ors

in

Ornatin

Glu.Gly.Pro.Phe.Gly.Pro.Gly

Glu.Gly.Pro.Phe.Gly.Pro.Gly

Glum---.Pro.Met*Lys.Ala*Thr

Gl".--'.Pro.let.Lys.Ala.Thr

Thr.Gl"*Gly-As"*Thr

Ala

Val.ValrTyr.Thr'Asp

Cys Thr.---.---.---*Gl".Ser

Ile.Thr*Tyr*Thr.Asp

Cys Thr*---.---.---.Gl".Ser

Val.Val.Tyr*Thr*Asp

Cys Thr.---.---.---.Gl".Ser

P,.u.Gl".GlY.Ser.Ala

Cy.g

---.---.---.---.A*".---

n

pr".Gl".GlymSermAla

cys

___._---_--.---.A*".---

Pro*Glu.Gly*Met.Wet

Cys Ser.---.---'---.Arg.---

Pro.GlU=Gly.Met.Wet

Cys Ser.---*---'---'Are.---

Ser.Ala.Ala.Gl".Val

Cys ---*---.Leu.Lys.Gly*Lys

-Pru.Arg.Leu.Pro.Gl"

---v---.---.---.Gln*Gly

Ile~Thr~Val~Arg~Pro~Thr~Lys~Asp~Glu~Leu~Leu~Thr

Gly~Glu~Phe~Arg~Glu~La"

I t

Gly~Gl"~Gly.As"~Lys

Cys Ile*Le".Gly.Ser

Gly~Lys~Gly~Asn~LyS

Cys Ile.Le".Gly*Ser

Gly~Gl"*Gly~Asn~Lys

Cys Ile.Le".Gly.~

Arg.Val

rl

.---.liis.---

Cys Pro.Hls.Gly.Phe

Arg*Val.---.Tyr.---s&s

Ser.llis.Gly*Phe

Arg.Lys.---.Thr.---*Cys

Pro~Asn~Gly~Le"

Arg*Lys=---.Thr*---'.Cys

Pro.Asn.Gly.Leu

Arg.Met'---.Phe.---

Cys Lys.Phe.Gly.Phe

Arg.Ile*---*---.Arg,Cys~Lys*Tyr.Gly*Leu

P~o.P~U.G~Y.G~".---*CYS

Arg.Phe.Pro*&g

Thr.Val.Gly.---'ArerCys

As"*Phe.Ala.m

L

Cys Glu~Gly.Ser*As".Val

Cys

Cys Glu.Gly*Ser.Asn.Val

CYs

Cys Glu.Gly~Ser.As".Val

Cys

Cys ---.Ser.Gly.Val.Arg

Cys

Cys ---=Pro.Glu.Val*Arg

Cys

I i

Cys Lys.Ile~Asp'Ile.As"

Cys

Cys Lys.Ile~Asp*Ile.As"

CYS

Cys ---.As"=Lys~Ile~Gl"

Cys

CYS ---~As"~Gl"~Val~His

CyS

CYS ---~As"~Lys~Asp*Gl"

CyS

Cys ---.Asp.Gly.Lys.Pro

CYs

Hirudi"

1

Gly~Gl"~Asn~Le"~---.---

CyS Le"

Hirudln

2-Lys47

Gly*Gl"~Asn~Le"*---.---

CyS Leu

Hirudisi"

Gly.Gln.Asn.Leu.---*---

Cys Le"

Antistasln

domain

1

Ile*Ile.Thr*Asp.Arg.---

CYs Thr

Ghilante"

domain

1

Ile*Ile*Thr.Asp.Arg.---

CYs Thr

Antistasin

domain

2

---.Leu.Thr*Asn.Lys.---

Cys Asp

Ghilante"

domain

2

---.Leu.Thr*As".Lys.---

Cys Asp

Hydra domain 1

u

Gl"

Hirustasi"

___.___.___.___.___.___

Cys Val

Decorsi"

Asp~Asp~Gln~Glu~Lys~---

CyS Le"

ornatin

cl

Gly~Gln~Pro'Asp~Lys'LyS

CYS Arg

Hlrudi"

1

Asp*Gly~Glu~Lys~As"*Gln

Hirudin

2-Lys47

As".Gly.Lys~Gly*As"*Gl"

Hirudisin

Gl~~As~.Ser.Lys~As"~Gl"

Antistasin

domain

1

Gln.Arg.Ser*Arg*Tyr.Gly

Ghilanten

domain

1

Gl".Arg.Ser.Arg.Tyr.Gly

Antistasin

domain

2

Lys.Arg.Asp*Lys.Leu.Gly

Ghilanten

dosain

2

Lys.Arg.Asp.Lys*Le".Gly

Hydra domain 1

Gl".Gl".Asp~Gl"~As"*Gly

Hirustasi"

Lys~Ly~.Asp~Gl"~Asn~Gly

Decorsin

Gly.As~.Ala.Asp.Pre'TYr

ornatin

~'Asp.As".Asp~Asp.Lys

cys

CYS

CY

S CYS

CY

S

CYS

CYS

CYS

CYS

CY

S CYS

-

Ile

Hirudi"

1

Gl"*Ile.Pro~G1"~G1u.Tyr~Le"'Gl"

Hirudin

2-Lys47

Gl"~Ile.Pro~Gl"~Glu.Tyr~Le"*Gl"

Hirudisln

Gl"~Ile~Pro~Gl"~Glu~Tyr-Leu*Gl"

Fig

. 2.

Am

ino-

acid

se

quen

ces

of t

he

mos

t ab

unda

nt

hiru

din

iso-

inhi

bito

r va

rian

t 1

(hir

udin

l)

, of

the

hi

rudi

n is

o-in

hibi

tor

vari

ant

2 be

arin

g a

Lys

re

sidu

e at

pos

itio

n 47

(hi

rudi

n 2-

Lys

47),

of

bio

synt

heti

c hi

rudi

n (h

irud

isin

),

of a

ntis

tasi

n is

o-in

hibi

tor

A

dom

ains

1

and

2, o

f gh

ilant

en

dom

ains

1

and

2,

of

Hyd

ra-a

ntis

tasi

n do

mai

n 1

(Hyd

ra

dom

ain

l),

of

hiru

stas

in,

of

deco

rsin

an

d or

nati

n is

o-fo

rm

E

(Ryd

el

et a

l.,

1991

; D

unw

iddi

e et

al

., 19

93;

Kre

zel

et a

l.,

1994

; So

lIne

r et

al

., 19

94).

T

he

arro

w

mar

ks

the

Pt-

P,’

re

acti

ve

site

bo

nd

in

anti

stas

in

dom

ain

1,

ghila

nten

do

mai

n 1,

Hyd

ra-a

ntis

tasi

n do

mai

n 1

and

hiru

stas

in,

inhi

biti

ng

the

Fac

tor

Xa

and

tiss

ue

kalli

krei

n ac

tivi

ty.

The

A

rg-G

ly-A

sp

sequ

ence

in

hi

rudi

sin,

de

cors

in

and

orna

tin,

an

tago

nist

s of

gl

ycop

rote

in

IIb-

IIIa

as

w

ell

as

inhi

bito

rs

of

plat

elet

ag

greg

atio

n,

has

been

un

derl

ined

. C

ys

resi

dues

ar

e bo

xed.

G

aps

are

indi

cate

d by

da

shes

. M

ulti

ple

alig

nmen

ts

wer

e pr

oduc

ed

usin

g th

e G

enet

ics

Com

pute

r G

roup

se

quen

ce

anal

ysis

so

ftw

are

pack

age

(GC

G

vers

ion

7.

1)

usin

g a

Vax

/VM

S sy

stem

(D

ever

eux

et a

l.,

1984

).


Cys22-Cys39), whereas the Gln49-Gln65 segment (Glu49-Gln65 in hirudin 2-Lys47) is free of intramolecular covalent cross-links and rich in negatively-charged Asp and Glu residues. A sulfated Tyr residue is present at position 63 close to the C-terminus of hirudin (see Figs 2 and 3) (Dodt et al., 1985; Seemiiller el al., 1986; Mao et al., 1987; Krstenansky et al., 1988a; Scharf et al., 1989).

The disulfide-looped structure shown in Figs 2 and 3 is characteristic of all natural and synthetic hirudin variants (Seemtiller et al., 1986; Mao et al., 1987; Krstenansky et al., 1988a; Tripier, 1988; Scharf et al., 1989; Krezel et al., 1994), of Factor Xa and tissue kallikrein inhibitors from the leeches Huementuriu oficinulis (antistasin), Huementuriu ghiliunii (ghilanten) and Hirudo medicinalis (hirustasin), of the antistasin- like six-domain repeat present in the inhibitor from the primitive metazoan Hydra, of the antagonists of glycoprotein IIb-IIIa as well as of the platelet aggregation inhibitors from the leeches Mucrobdellu decoru (decorsin) and Plucobdellu ornutu (ornatin) (Krezel et al., 1994; Sollner et al., 1994). Moreover, the disulfide bridge pattern of hirudin (Cys-X,,,-C~S-X-C~~-X~C~S-X-C~~-X~_~~-C~S) (Krezel et al., 1994) is topologically reminiscent (although in reverse order) of that observed in the epidermal growth factor domains present in several serine proteinases and extracellular proteins (Appella et al., 1988). Nevertheless, hirudin and the related inhibitors mentioned above do not bear any sequence homology to the epidermal growth factor domains (Appella et al., 1988; Krezel et al., 1994).

At variance with the homologous serine proteinase inhibitors antistasin, ghilanten and hirustasin (see Chapter 4), hirudin does not obey the ‘standard mechanism’ of serine proteinase inhibition. In fact, hirudin does not fill thrombin active site region with a reactive site loop adopting the canonical substrate-like conformation (Laskowski and Kato, 1980; Amiconi et al., 1987; Bolognesi et al., 1987; Bode and Huber, 1991, 1992; Ayala and Di Cera, 1994; Schreuder et al., 1994; Stubbs and Bode, 1994) (see Chapter 2, Fig. 1 and below).

l4

163

R

‘i;$ OH 163

Fig. 3. Stereo view of the hirudin 2-Lys47 molecule as observed in the binary human a- thrombin-inhibitor complex (Rydel et al., 1991). Note the extended hirudin 2-Lys47 C-terminal polypeptide chain which binds to the fibrinogen exosite of the serine proteinase. The inhibitor

residues are identified by the suffix ‘I’.

Proteinase Inhibitors from Leech Hifudo medicinalis 235

The three-dimensional structures of free native and recombinant hirudin forms have been investigated in solution, by NMR techniques (Folkers et al., 1989; Haruyama and Wtithrich, 1989; Szyperski et al., 1992a, b), and in complexes with thrombin, by X-ray crystallography (Griitter et al., 1990; Rydel et al., 1990, 1991; Bode and Huber, 1991, 1992; Bode et al., 1992a, b; Vitali et al., 1992; Priestle et al., 1993). Moreover, the structure of the binary complexes of hirugen (corresponding to the iV-acetyl-Asn53-Leu64 C-terminal segment of hirudin), hirulog 1 (corresponding to the (D)Phe-Pr~Arg-Pro-(Gly),-Asn53-(desulfot~Tyr63)-Leu64 C-terminal segment of hirudin), hirulog 3 (identical to hirulog 1, but bearing a @home-Arg residue as the third residue), hirutonin 2 (corresponding to the N-acetyl-(D)Phe-ProArg- (CH,),-C(O)-Asn49-Gln65 C-terminal segment of hirudin), hirutonin 6 (corresponding to the N-acetyl-(D)Phe.Pro.Arg-(CH,),-C(O)-(Gly)~-Asp55-Leu64 C-terminal segment of hirudin) with thrombin have been solved (Skrzypczak-Jankun et al., 1991; Ni ef al., 1990; Qiu ef al., 1992, 1993; Zdanov et al., 1993; Vijayalakshmi et al., 1994).

In free hirudin 1, the N-terminal domain shows a compact disulfide-linked globular structure, whereas conformational disorder is observed for residues Vall (Ilel in hirudin 2-Lys47), Va12 (Thd in hirudin 2-Lys47), Gly31-Lys36 (Gly31-Gly36 in hirudin 2-Lys47), and for the whole Pro48-Gln65 C-terminal segment (Folkers et al., 1989; Haruyama and Wtithrich, 1989; Szyperski et al., 1992a, b). The inhibitor N-terminal domain maintains a compact conformation also upon binding to thrombin (Griitter et al., 1990; Rydel et al., 1991; Priestle et al., 1993). Alternatively, the C-terminal polypeptide segment, which is flexible in free hirudin 1, becomes ordered as a result of extensive se&e proteinase-inhibitor interactions, the polypeptide chain being stabilized in a rather extended conformation (see Fig. 3) (Griitter et al., 1990; Rydel et al., 1990, 1991; Bode and Huber, 1991, 1992; Bode et al., 1992a, b; Vitali et al., 1992; Priestle et al., 1993).

The peculiar tertiary structure of hirudin (N-terminal globular- and C-terminal extended- regions; see Fig. 3) reflects the tight cystine cross-linking present in the N-terminal region (see Fig. 3), whose structural organization is based on two pairs of antiparallel P-strands. The first pair of hirudin 1 is rather short and built up by the Cysl4-Cysl6 and by the Asn2&Cys22 P-strands. The second pair is more extended, comprising the antiparallel Lys27-Gly31 and Lys36-Va140 (Gly36-Va140 in hirudin 2-Lys47) P-strands (see Figs 3 and 4). Due to the presence of the three disulfide bridges in the N-terminal region of hirudin 1, the four antiparallel p-strands are constrained in their mutual orientations, and cross-linked (through the Cys6-Cysl4 disulfide bridge) to the Vall-Leu13 segment (Ilel-Leu13 in hirudin 2-Lys47). The hirudin 1 Thr4-Leu13 region adopts the conformation of a rather wide loop, devoid of regular secondary structure. In addition, from the pH-dependence of the lH NMR line shifts corresponding to different charged residues of the recombinant hirudin 1 as well as of its mutants, the existence of a series of transient H-bonds is suggested. These involve the backbone amide protons of several residues (such as Glu17-Cys39, Ser32-Glu35, Gly25-Glu43 and Asp33-Glu35), with pK values ranging between 3.7 and 4.7, indicating that a network of weak interactions also contributes to keep the core of the hirudin molecule in a given structural arrangement (Szyperski et al., 1994). Despite this interlaced polypeptide organization, the presence of Gly residues in the reverse turn regions connecting the four p-strands may account for the local flexibility observed for this portion of hirudin


Fig. 4. Stereo view of the Ca backbone of the human a-thrombin-hirudin 2-Lys47 complex (Rydel et al., 1991). The hirudin 2-Lys47 molecule is highlighted in thick bonds.

in solution (Folkers et al., 1989; Haruyama and Wiithrich, 1989; Rydel et al., 1990, 1991; Szyperski et al., 1992a, b; Bode ef al., 1992a).

The extended conformation adopted by the C-terminal polypeptide chain of hirudin 1 and of hirudin 2-Lys47, once bound to thrombin, is compatible with its thorough solvent exposure when the inhibitor (hirudin 1) is free in solution. The polypeptide chain spanning from Pro46 to Gln65 can be schematically divided into two segments, which merge around residue Gly54, where the crystallographic electron density maps indicate a certain degree of structural flexibility (Bode et al., 1992a). The first part of the C-terminal region, from Pro46 to His51, adopts a polyproline II left-handed helical conformation, and spans about 18 A. The second segment (Asp55-Pro60 residues, about 16 8, long) is followed by a type 3,, helical turn (Glu61-Leu64 residues) (see Fig. 3) (Folkers et al., 1989; Griitter et al., 1990; Rydel et al., 1991; Bode et al., 1992a).

The thrombin-hirudin contact area (1800 Az) is significantly larger than those commonly observed for canonical serine proteinase-protein inhibitor systems (a contact area of 714 A2 has been reported for the bovine ol-chymotrypsin-eglin c complex; see Chapters 2 and 5. Frigerio et al., 1992), comprising 33 thrombin and 26 hirudin residues, for a total of 216 interatomic contacts shorter than 4.0 8, (Grtitter et al., 1990; Rydel et al., 1990, 1991; Bode et al., 1992a).

In the thrombin-hirudin 1 and thrombin-hirudin 2-Lys47 complexes, three selected contact regions can be distinguished, on the inhibitor side: (1) the N-terminal Vall-Tyd segment (Ilel-Tyr3 in hirudin 2-Lys47); (2) the N-terminal disulfide-linked globular domain; and (3) the negatively-charged C-terminal elongated segment (see Figs 3 and 4). Along these regions, hirudin embraces thrombin, starting from the proteinase active site cleft and extending for approximately 35 A across the enzyme surface through the fibrinogen binding exosite (see Fig. 4) (Grutter et al., 1990; Rydel et al., 1990, 1991; Stubbs and Bode, 1993b, 1994).

The hirudin 1 Vall-Tyr3 N-terminal residues (Ilel-Tyr3 in hirudin 2-Lys47) form a short parallel p-sheet structure with the thrombin Ser214-Gly219 segment, close to the


enzyme catalytic triad. Alternatively, an antiparallel interaction. is observed between substrate-like serine proteinase protein inhibitors (e.g. eglin c) and the same segment of the cognate enzyme (e.g. subsites S, and S, of bovine o-chymotrypsin) (see Chapter 2, Fig. 1, Chapter 5, Fig. 9) (Huber and Bode, 1978; Laskowski and Kato, 1980; Read and James, 1986; Amiconi et al., 1987; Bolognesi et al., 1987; Bode and Huber, 1991, 1992; Frigerio et al., 1992).

The Vall free N-terminal group of hirudin 1 (Ilel in hirudin 2-Lys47) is hydrogen bonded to the thrombin active site Ser195 OG atom (the nucleophile species during catalysis) and to the enzyme Ser214 0 atom. Moreover, the inhibitor Vall and Tyr3 side-chains (Ilel and Tyr3 in hirudin 2-Lys47) are located in two apolar pockets on the thrombin surface. Thus, Vall (Ilel in hirudin 2-Lys47) falls in the serine proteinase S2 recognition subsite, whereas Tyr3 occupies the so called ‘aryl binding site’. The primary specificity subsite of thrombin (S,) is only marginally engaged by the side-chain of the hirudin 1 Va12 residue (Thd in hirudin 2-Lys47), which does not fill the S, pocket and does not establish a salt link with Asp189, as observed commonly for synthetic and natural inhibitors and for substrates of trypsin-like serine proteinases bearing Arg or Lys residues at their P, position (see Figs 4,5, Chapter 2, Fig. 1, Chapter 5, Fig. 9). Indeed, hirudin binding to thrombin is not prevented by the enzyme inactivation through small synthetic inhibitors (see next section) (Huber and Bode, 1978; Laskowski and Kato, 1980; Read and James, 1986; Stone et al., 1987; Amiconi et al., 1987; Bolognesi et al., 1987; Bode et al., 1989, 1992a, b; Grtitter et al., 1990; Rydel et al., 1990, 1991; Bode and Huber, 1991, 1992; Stubbs and Bode, 1993b, 1994; Schreuder et al., 1994; Stone, 1995).

The N-terminal disulfide-linked globular domain of hirudin 1 and hirudin 2-Lys47 is contacting thrombin at a limited extent through two ion pairs (AspS-Arg221A and Glu17-Arg173, inhibitor-proteinase residues, respectively) and two hydrogen bonds

Fig. 5. Stereo view showing the region surrounding the primary specificity subsite (St) of human u-thrombin in the presence of the hirudin 2-Lys47 Ilel and Thd N-terminal residues (Rydel et al., 1991). The alcoholic oxygen atom of hirudin Thr2 is 7.1 8, away from the human a-thrombin Asp189 carboxylate, at the dead end of the enzyme S, subsite. The enzyme and the inhibitor

residues are identified by suffix the ‘H’ and ‘I’, respectively.


(Serl9 G-Glu217 OEl and Va121 N-Glu217 OEl, inhibitor-proteinase residues, respectively), reminiscent of protein-protein aspecific surface contacts. In addition, van der Waals and weak polar interactions, often mediated by solvent water molecules, occur in this region of the serine proteinase-inhibitor complex. Thus, the N-terminal disulfide-linked globular domain of hirudin 1 and hirudin 2-Lys47 is in weak contact with the thrombin surface, nevertheless preventing the accessibility of macromolecular inhibitors and substrates to the enzyme active centre (see Fig. 4) (Bode et al., 1992a; Stubbs and Bode, 1993b; Betz et al., 1991a, 1994).

The negatively-charged C-terminal elongated segment of hirudin 1 and hirudin 2-Lys47 finds structural and (a)polar counterparts in the thrombin fibrinogen-binding exosite, a groove on the serine proteinase surface rich in arginyl and lysyl residues, but also displaying apolar patches. In this serine proteinase-inhibitor contact region, 116 (a)polar intermolecular interactions (shorter than 4.0 A) occur (Bode et al., 1992a; Stubbs and Bode, 1993b). In this respect, the hirudin 1 Gln49 (Glu49 in hirudin 2-Lys47), Ser50, His51, Asp55, Glu57 and Glu58 residues contact the thrombin Lys60F, Glu39, Arg73, Lys149E, Arg75, Arg77A and the C-terminal Lys36 side-chains, respectively (see Fig. 4). Moreover, the inhibitor Phe56, Ile.59, Pro60, Tyr63 and Leu64 residues provide extended hydrophobic contacts with selected apolar patches present in the thrombin fibrinogen-binding exosite (Phe34, Leu65, Tyr76 and Ile82) (Grutter et al., 1990; Rydel et al., 1990, 1991; Bode et al., 1992a; Stubbs and Bode, 1993b, 1994).

The hirudin Tyr63 residue may contact thrombin through different geometries. Although in the crystal structure of the binary complex of thrombin with the hirudin 2-Lys47, the Tyr63 residue is not sulfated, the enzyme Lys81, LyslO9 and LysllO side-chains fall close to the potential sulfate site in the binary complex. Such finding suggests that direct salt-link interaction(s) may occur between the enzyme residues and the hirudin Tyr63 sulfated side-chain (Rydel et al., 1991; Bode et al., 1992a; Stubbs and Bode, 1993b). However, in the crystal structure of the binary complex of thrombin with hirugen, bearing a sulfated Tyr63 residue, a direct serine proteinase-inhibitor ion pair is not observed. In fact, the inhibitor Tyr63 sulfate group is linked to thrombin hydrogen-bond donors and to one solvent water molecule (Skrzypczak-Jankun et al., 1991; Vijayalakshmi et al., 1994).

Inspection of the thrombin and hirudin amino-acid residues facing along the serine proteinase fibrinogen-binding exosite demonstrates a remarkable linear distribution of complementary electrostatic charges. However, not all of the acidic residues of the hirudin tail are capable of direct interaction(s) with the serine proteinase exosite, due to steric reasons or unfavourable side-chain orientations. Simulation of thrombin and hirudin electrostatic fields demonstrates that, even in the absence of direct interactions, the facing serine proteinase and inhibitor charged residues do provide intermolecular complex stabilization through the establishment of electrostatic fields, on the two proteins, which are remarkably complementary in shape and sign (Karshikov et al., 1992). Thrombin/hirudin electrostatic charge complementarity plays a role not only in the stabilization of the enzyme-inhibitor binary complex, but also in guiding the pre-orientation of the binding surfaces, facilitating association of the complex, a process known to be very fast (Dodt et al., 1988). In particular, studies of complexes formed by thrombin and hirudin C-terminal fragments (i.e. residues 52-65) by use of one- and


two-dimensional NMR techniques (Ni et al., 1990) suggest that the distribution of polar and apolar residues in this region allows optimization of binding to the exosite of thrombin. The potential role of different residues has been established clearly by the X-ray structure of the complex of thrombin with peptide inhibitors, which mimic the C-terminus of hirudin, and interact with the exosite of thrombin (Ni et al., 1990; Qiu et al., 1992, 1993).

Functional Aspects

Hirudin is the most potent and selective natural inhibitor of thrombin. No significant inhibitory cross-reaction with other serine proteinases is known (Haycraft, 1884; Markwardt, 1955, 1986, 1988; Seemiiller ef al., 1977, 1986; Schnebli and Braun, 1986; Fenton, 1989; Ascenzi et al., 1992a; Stone and Maraganore, 1993; Stubbs and Bode, 1993a, b, 1994; Banner et al., 1994; Stone, 1995; Verlinde and Hol, 1994). The different natural hirudin forms show the same affinity (i.e. K values) for thrombin (Scharf et al., 1989). However, values of K for recombinant hirudin (displaying desulfated-Tyr63) are lower by about one order of magnitude than those of the native inhibitor (see Tables 1 and 2) (Stone and Hofsteenge, 1986; Braun et af., 1988; Ascenzi ef al., 1992a). The interaction of human a-thrombin with hirudin is characterized by a strikingly high affinity (K= 8.3~1013 M-l, at pH 7.5 and 21”(Z), which is only scarcely affected by temperature (Ascenzi et al., 1992a). The high value of K finds its dynamic background in a very fast, diffusion-limited, association rate constant (k,, = 1.1 X 109M-1sec-1) and a very slow dissociation rate constant (k,, = 1.3 x lo-%ec-I), both contributing to yield a virtually irreversible enzyme- inhibitor binary complex (Ascenzi et al., 1992a).

A three-step mechanism for molecular recognition of hirudin by thrombin has been proposed (Jackman et al., 1992; Ayala and Di Cera, 1994). Initially, the C-terminal acidic tail of hirudin binds to the fibrinogen-binding exosite of thrombin, displacing one chloride ion from the serine proteinase surface. This event triggers a conformational transition of thrombin, increasing the inhibitor accessibility to the enzyme catalytic pocket. Finally, the compact N-terminal domain of hirudin is accomodated in the thrombin region surrounding the active centre.

Table 1. Values of thermodynamic and kinetic parameters for the binding of hirudin and of the bovine basic pancreatic trypsin inhibitor (BPTI) to human a-, p- and y-thrombin, at pH 7.5 and

21 .o”c*

Proteinase Inhibitor (5) k (M-1 .L)

k off (set-1)

a-Thrombin P-Thrombin y-Thrombin a-Thrombin P-Thrombin y-Thrombin

Hirudin Hirudin Hirudin BPTI BPTI BPTI

8.3 x 10’3 5.6 x 10” 7.9 x 10’ 1.2 x 103

2.5 x 103 9.1 x 103

1.1 x 109 1.3 x 10-s 1.7 x 10’ 3.1 X 1rY 1.3 x 104 1.6 x lo” 4.0 x 104 7.0 x 10’ 5.0 x 104 3.0 x 10’ 8.5 x 104 1.0 x 10’

*From Ascenzi et al. (1992a).


The affinity of hirudin for human thrombin is dramatically reduced whenever the serine proteinase p- and/or y-loops are removed, such as in the p- and y-thrombin, respectively, clearly indicating that these thrombin regions are crucial for the formation of the enzyme-inhibitor complex (see Ta.ble 1) (Stone et al., 1987; Ascenzi et al., 1992a). This behaviour can be almost totally attributed to a corresponding marked decrease of the kinetic association rate constant, which is accompanied by only a moderate increase of the dissociation rate constant (see Table 1) (Ascenzi et al., 1992a). The specific role of the p- and y-loop(s) in the recognition mechanism of thrombin-hirudin interaction is further outlined by the observation that cleavage of the enzyme polypeptide sequence in different region(s), such as in @,-- and a-thrombin (see above), does not induce any appreciable variation of the equilibrium affinity constant (Stone et al., 1987; Ascenzi et al., 1992a).

The interaction mechanism between thrombin and hirudin displays features which are remarkably different from those observed for the formation of the enzyme-inhibitor complex in other serine proteinases. Thus, from the functional viewpoint, very efficient protein inhibitors of trypsin and chymotrypsin, such as the bovine basic pancreatic trypsin inhibitor (Kunitz-type inhibitor; BPTI), are poor inhibitors of thrombin. Such a consideration is strengthened further by the fact that the affinity trend for BPTI interaction with human thrombin is the opposite of what observed for hirudin, the affinity of a-thrombin for BPTI being smaller than that of y-thrombin (see Table 1) (Ascenzi et al., 1988). Altogether, the opposite effect exerted by the thrombin l3- and y-loop(s) on the modulation of the affinity for hirudin and BPTI suggests indeed that in the case of hirudin, they act as anchoring element(s) for the stabilization of the complex, whereas in the case of BPTI, they represent a steric limitation to the adduct formation (Ascenzi et al., 1992a).

The interaction of hirudin with thrombin displays a pH-dependence for the equilibrium constant showing the same qualitative features for W, p- and y-thrombin, even though the absolute affinity constant values are markedly different (see Table 1) (Ascenzi et al., 1992a). Thus, for all the three forms of thrombin, the pH-dependence of K shows a bell-shaped pattern which has to be attributed to the perturbation of the acid-base equilibrium of (at least) two groups, upon formation of the binary complex. A completely different pH-dependence curve is obtained in the case of BPTI interaction with human y-thrombin, which can be described by the perturbation of the protonation of a single group (Ascenzi et al., 1992a), further supporting the view that the two macromolecular inhibitors bind to thrombin in drastically different fashions (Bode and Huber, 1991, 1992; Ascenzi et al., 1992a).

The bell-shaped pH-dependence of the equilibrium constant for the binding of dative hirudin is also observed for the association of the recombinant inhibitor, in which, from the kinetic standpoint, the pH-dependence can be attributed wholly to the proton-linked effect on the dissociation rate constant (Betz et al., 1992). The extension of the investigation on the pH effect to hirudin mutants allowed the determination of the requirement of three ionization groups (the N-terminal of the inhibitor, and the N-terminal Ile16 and His57 of the serine proteinase) for the full description of the phenomenon. In fact, the acetylation of the N-terminus of recombinant hirudin causes the disappearance of one protonation event (Betz et al., 1992). Alternatively,


no other mutation on the hirudin molecule, such as: (1) the removal of the last 13 residues (containing several negatively-charged side-chains); (2) substitution by uncharged residues of all lysyl side-chains at positions 27, 36 and 47; or (3) substitution of His51 by Gin, induces any variation in the pH-dependence of thermodynamic and kinetic parameters for the formation of the thrombin-inhibitor complex, ruling out a role of these portions of the molecule on the observed behaviour (Betz et al., 1992). Thus, thrombin ionizing residues affecting hirudin binding (i.e. Ile16 and His57) are not involved neither in the p- or in the r-loop (Ascenzi et al., 1992a).

In spite of the fact that thrombin demonstrates a catalytic competence for substrates displaying positively-charged residues at the P, position (such as lysyl and arginyl side-chains), the functional role of lysyl residues in the thrombin-hirudin interaction appears very minor. Thus, mutations involving the disappearance of the positive charge on either Lys27 or Lys36 have no effect on the affinity constant; alternatively, in the case of Lys47 only a 9-fold decrease for K is observed (see Table 2) (Braun et al.,

1988). Such an observation indeed supports the view that the Lys47 residue of hirudin is positioned near the thrombin recognition centre, but the very limited functional effect observed upon neutralization of this residue clearly indicates that, unlike for other macromolecular inhibitors (such as BPTI or Kazal-type inhibitors, see Bode and Huber, 1991, 1992), this ionic interaction is not functionally crucial for the formation of the thrombin-hirudin complex.

Even though lysyl residues appear to play a very minor role in the serine proteinase- inhibitor interaction, the binding of hirudin to thrombin is indeed modulated by electrostatic interactions, as indicated by the fact that the equilibrium affinity constant, besides the proton-linked effect, is significantly affected by ionic strength (Stone et al.,

1989) and by specific ions (De Cristofaro et al., 1992). Thus, the increase of ionic strength appears to induce an affinity decrease, which mostly results from a slower kinetic association rate constant (Stone and Hofsteenge, 1986). The structural basis of this behaviour have been elucidated to a remarkable extent by functional studies employing site-directed mutants of hirudin, outlining the primary role played by the cluster of negatively-charged residues present in the C-terminal region of the hirudin molecule (see Table 2) (Braun et al., 1988; Betz et al., 1991a; Karshikov et a/. , 1992). In particular, it has been observed that the effect can be mostly attributed to Glu residues located at positions 57, 58, 61 and 62, since their substitution by Gln brings about an affinity decrease, with different contributions for the different amino-acid residues (see Table 2) (Braun et al., 1988; Betz et al., 1991a). In fact, individual substitutions of the four Glu side-chains induce a much bigger effect for Glu57 and, to a somewhat lesser extent, for Glu58 than for Glu61 and Glu62 (see Table 2) (Betz et al., 1991a). However, when double, triple and quadruple mutants (all employing glutamine) have been obtained for these negative residues of hirudin, the interpretation appears much more complicated. Thus, while neutralization of Glu61 and Glu62 has an additive linear effect (indicating that each one contributes independently to the affinity constant), in the case of Glu57 and Glu58, which play a major role, the effect turns out to be coupled, as if neutralization of either one separately brought about a tertiary conformational change verifying the electrostatic contribution of the other one (see Table 2) (Betz et al., 1991a). Such a behaviour finds at least partial support in the X-ray crystal structure of the human a-thrombin-hirudin 1 and a-thrombin-hirudin 2-Lys47


Table 2. Values of the association equilibrium constant for the binding -of hirudin mutants with polar residues in the C-terminal region to human cu-thrombin

Hirudin form K (M-l)

Native* Desulfato-native* Recombinant (desulfato)t Lys36 --, Glnt Lys47 -+ Glnt Pro48 -+ Ala+ His51 -+ Glnt Asp53 + Ala§ Asp55 + Asn§ Glu57 + GlnO Glu58 -+ Gln§ Glu61+ Gln§ Glu62 + Glnt Glu57 + Gln-Glu58 + Glnt Glu61 --, Gln-Glu62 + Gln§ Glu57 -+ Gln-Glu58 + Gln-

Glu62 + Glnt Glu57 + Gln-Glu58 + Gln-

Glu61 -+ Gln-Glu62

5.9 x 10’3

4.8 x 10’2 4.3 x 10’2 4.5 x 10’2 5.0 x 10” 7.7 x 10” 4.2 x 10’2 4.8 x 1012 1.7 x 10’2 4.3 x 10” 5.6 x 10” 2.7 x 1012 1.8 x 1012 4.2 x 10” 1.1 x 10’2

1.2 x 10”

+ Gin? 8.0 x 10’0

*pH = 7.8 and 37.O”C, from Stone and Hofsteenge (1986); tpH = 7.8 and 37.0°C, from Braun et al. (1988); $pH = 8.3 and 25.O”C, from Dodt et al. (1990); $pH = 7.8 and 37.O”C, from Betz et al. (1991a).

complexes (Griitter et al., 1990; Rydel et al., 1990, 1991). In fact, hirudin Glu61 and Glu62, in spite of their contiguity in the primary structure, are oriented differently, the two carboxylate groups being well separated (by about 10 A). Alternatively, Glu57 and Glu58 are much closer, interacting. with Arg75 and Arg77 of human ol-thrombin, respectively (Griitter et al., 1990; Rydel et al., 1990), which might explain both their major role in influencing individually the affinity constant as well as their mutual interaction (Betz et al., 1991a). However, it must be remarked that a quantitative analysis based on the crystal structure of the electrostatic contributions arising from different negatively-charged residues failed to account for the observed effects on the thermodynamic equilibrium constant, since an unacceptably large discrepancy has been observed between the actual and expected role of Asp55 (see Table 2) (Karshikov et al., 1992). Thus, this residue, which, according to the crystal structure, should be interacting with Lys149 and Arg73 of human ol-thrombin (Rydel et al., 1990), does not seem to play any role and its neutralization does not bring any effect, opening the question on the possibility of some difference between the crystal and solution structural arrangement of the thrombin-hirudin complex (Betz et al., 1991a).

The importance of the conformation of the C-terminal portion of hirudin on the affinity constant for the serine proteinase-inhibitor complex formation is also strengthened by the observations of the functional effect exerted by mutations of non-polar residues

Proteinase Inhibitors from Leech Hitudo medicinalis 243

in this region of the molecule (see Table 3) (Betz et al., 1991b). Thus, in the case of Phe56, a large functional effect is observed when it is substituted by a non-polar branched amino-acid residue, such as Leu, Ile, Val or Thr, whereas only minor or no effects are detected if the mutation deals with Ala, Trp or Tyr (see Table 3). A similar qualitative effect is also observed for mutations involving hirudin Pro60 and Tyr63 residues (see Table 3) (Betz et al., 1991b). Such observations appear in line with the reported structural evidence of very important interactions between apolar residues present in the C-terminal region of hirudin and the exosite cleft of human ol-thrombin (Qiu et al., 1992, 1993).

Sulfation of Tyr63 (a residue located very close to the C-terminal of hirudin) in the native inhibitor is important in characterizing the affinity constant for thrombin, since its desulfation leads to a decrease of approximately lo-fold for the equilibrium association constant (see Table 2) (Stone and Hofsteenge, 1986). Nitration or iodination of Tyr63, which is desulfated in recombinant hirudin, almost restores the native inhibitor affinity for human cx-thrombin (Winant et al., 1991), indicating the importance of this ionic interaction for the functional role played by Tyr63, as also suggested by X-ray crystal structures (Stubbs and Bode, 1994). Although very important, the interaction of the C-terminal region of hirudin is not the only determinant in thrombin inhibition, as indicated by the fact that the polypeptide formed by the last 21 residues of hirudin (i.e. the 45-65 peptide) shows decreased affinity (by approximately 2000-fold) with respect

Table 3. Values of the association equilibrium constant for the binding of hirudin mutants with apolar residues in the C-terminal region to human a-thrombin

Hirudin form K (M-‘)

Native* 5.9 x 1013

Recombinant (desulfato)t 4.3 x 1012 Phe56 + Tyr$ 4.8 x 10’2 Phe56 + Trpf 2.5 x 10’2 Phe56 -+ Ala+ 2.1 x 10’2 Phe56 -+ Leu$ 1.4 x 10” Phe56 + Ilet 1.4 x 10” Phe56 + Val$ 7.3 x 1010 Phe56 + Thr$ 3.8 x 10’0 Pro60 -+ AlaS 4.8 x 10” Pro60 -+ Gly$ 3.3 x 10” Tyr63 --+ Phet: 3.8 x 10’2 Tyr63 -+ Ala* 1.9 x 10’2 Tyr63 + Glu$ 2.1 x 10’2 Tyr63 + Leu$ 1.3 x 1012 Tyr63 -+ Val$ 6.0 x 10” Phe56 -+ Val-Asp55 + Asn$ 2.8 x 10’0 Phe56 -+ Val-Glu57 --+ Gln$ 9.9 x 109

*pH = 7.8 and 37.0°C, from Stone and Hofsteenge (1986); tpH = 7.8 and 37.O”C, from Braun et al. (1988); $pH = 7.8 and 37.O”C, from Betz et al. (1991b).


to the whole inhibitor (Schmitz et al., 1991). This peptide binds only to the exosite cleft of human cw-thrombin, with an affinity constant of 1.3X106 M-l, which is significantly higher than that observed for the peptide formed by the last 11 residues of hirudin (i.e. the 55-65 peptide; K = 2.7~ 105 M-l) (Krstenansky and Mao, 1987; Krstenansky et al., 1988b). Alternatively, the l-47 hirudin peptide, from which the C-terminal region had been removed, displays a slightly higher affinity with (K = 2.5~106 M-r) (Schmitz et al., 1991), and the 143 hirudin peptide loses thrombin inhibition properties almost completely (Chang, 1983).

Such evidence is supported by observations on hirudin mutants involving the N-terminal amino-acid residues. Thus, the elimination of the positive charge of the N-terminus by acetylation of the a-amino group brings about a 5000-fold decrease for the association equilibrium constant (Wallace et al., 1989). Furthermore, replacement of the first two Val residues in hirudin 1 by Phe does not produce any appreciable effect; alternatively, substitution by Leu leads to a 50-fold affinity decrease, the effect being totally attributed to residue Va12 (see Table 4) (Wallace et al., 1989). Even larger effects have been observed if either one or both Val side-chains were substituted by charged residues, the effect being larger for Glu than for Lys (see Table 4). This behaviour is clearly reflecting the importance of hydrophobic interactions on the binding of the N-terminal region of hirudin to human cw-thrombin, and in particular, the close proximity of the first two residues (i.e. Vall-Va12 in hirudin 1 and Ilel-Thr2’in hirudin 2-Lys47) to the active centre of the enzyme (Bode et al., 1992a). Moreover, Tyr3 seems also to play a relevant role in the human a-thrombin- hirudin complex formation, since its substitution by

Table 4. Values of the association equilibrium constant for the binding of hirudin mutants in the N-terminal region to human a-thrombin

Hirudin form K (M-l)

Native* 5.9 x 1013

Recombinant (desulfato)t 4.3 x 1012

Vall + LeuS 4.3 x 10’2

Va12 + Leu$ 9.7 x 10’0

Vall + Glu$ 3.4 x 109

Va12 + Glu$ 4.0 x 109

Vall + Ile-Va12 + Ile$ 1.0 x 10’3

Vall + Phe-Va12 + PheS 4.2 x 10’2

Vail + Leu-Val2 -+ LeuS 1.0 x 10”

Vall -+ Ser-Va12 -+ SerS 5.7 x 109

Vall + Lys-Va12 + Lys$ 6.6 x 109

Vall + Gl;-Va12 -+ Glyf 1.4 x 109

Vall + Glu-Val2 + Glu$ 1.5 x 107

Tyr3 -+ Phe§ 7.9 x 10’2

Tyr3 + Trp§ 6.0 x 10’2

*pH = 7.8 and 37.O”C, from Stone and Hofsteenge (1986); tpH = 7.8 and 37.O”C from Braun et al. (1988); fpH = 7.8 and 37.0°C, from Wallace et al. (1989); §pH = 7.8 and 37.0% from Winant et al. (1991).

Proteinase Inhibitors from Leach Hirudo medicinalis 245

Thr leads a 500-fold decrease in the affinity constant (Lazar et al., 1991). However, although Tyr3 is conserved in all hirudin variants (Scharf et al., 1989), its substitution by hydrophobic amino-acid residues, such as Phe or Trp, enhances the inhibitor affinity for or-thrombin (see Table 4) (Winant et al., 1991).

A peculiar field of investigation, which has provided a great wealth of information, deals with the design and synthesis of peptides, which mimic the role of fragments of hirudin in order to accomplish only some of the coexistent actions exerted by the inhibitor (see Stone and Maraganore, 1993). Particular importance in this respect is played by synthetic peptides which bind only to the fibrinogen-binding exosite (leaving thus accessibility to the active ‘amidolytic’ centre), mirroring the interaction of the C-terminal portion of hirudin with human a-thrombin (Dennis et al., 1990). The main representative of this class is hirugen, which associates to human a-thrombin (Skrzypczak-Jankun et al., 1991) with a geometry closely similar to that reported for the last 11 amino acids of the human cw-thrombin-hirudin complexes (Griitter et al., 1990; Rydel et al., 1990, 1991; Bode et al., 1992a). Thus, hirugen behaves as a purely competitive inhibitor (with K = 1.9~106 M-l) for the human a-thrombin catalyzed release of fibrinopeptide A from fibrinogen, while the effect on the binding of antithrombin III is very contained (only a 2.5fold decrease of its affinity for ol-thrombin) and almost negligible on the hydrolysis of a synthetic substrate (Naski et al., 1990). A closely similar effect has been reported for the 54-65 hirudin peptide which shows a value of K of 1.4~ 106~-1 for human a-thrombin (De Cristofaro et al., 1993).

These peptides, although not able to block the catalytic site of human o-thrombin, are very efficient inhibitors of thrombin action on endothelial cells. Thus, they inhibit, in a concentration-dependent fashion, the thrombin-induced synthesis of prostaglandin I,, the platelet-activating factor, and the acquisition of an adhesive surface for leukocytes, occurring upon thrombin activation. These findings indicate that the exosite (the fibrinogen-binding cleft) of human cw-thrombin plays a central role in the interaction mechanism with platelet membrane receptor. Thus, the catalytic site of human (Y- thrombin plays a minor role in platelet activation (Prescott et al., 1990).

Bivalent peptides (hirulogs and hirutonins), which can interact both with the active site and the fibrinogen-binding exosite of human cw-thrombin have been designed (Skrzypczak-Jankun et al., 1991; Stone and Maraganore, 1993; Zdanov et al., 1993). Hirulogs bind with a very high affinity both to the human cx-thrombin ‘amidolytic’ catalytic site (K = 3.8~10*~-1) and the fibrinogen-binding exosite (K = 7.7xlO*~-i), whereas they display an almost unmeasurably low affinity for y-thrombin (Witting et al., 1992). However, the presence of an arginyl residue associating with the proteinase primary specificity subsite (S,) brings about a very slow, but appreciable cleavage of this inhibitor (k,,, = 1.OX1O-2 set-1); such a hydrolytic process disappears upon modification of Arg to @home-Arg (hirulog 3) (Kline et al., 1991). The formation of the human ol-thrombin-hirulog adduct is consistent with a four-step mechanism (Parry et al., 1994). The initial association of the C-terminal. region of hirulog (based on the hirudin amino-acid sequence) with the thrombin exosite is followed by a slow conformational change. This event precedes the interaction of the P, arginine residue of hirulog with the serine proteinase primary specificity subsite (S,) in a very rapid step; next, a further conformational change takes place. The value of the overall second-order association


rate constant for the human cu-thrombin-hirulog adduct formation (determined at low inhibitor concentration) is 4.0X 10s M-1 set-1, while the effective first-order dissociation rate constant is approximately 1 set-1.

A different class of thrombin inhibitors is based on the hirudin mutant, called hirudisin, in which residues Ser32, Asp33, Gly34 and Glu35 are replaced by Arg, Gly, Asp and Ser, respectively (see Fig. 2) (Knapp er al., 1992). Such a modification followed the observation of a very efficient inhibitory action of the human cw-thrombin- induced platelet aggregation and of the integrin-mediated attachment of endothelial and fibroblast cells to polystyrene walls exerted by a chimeric peptide, where the Arg-Gly-Asp sequence was coupled to the C-terminal fragment of hirudin (residue 53-64) (Church et al., 1991). The affinity of hirudisin for human cu-thrombin (K = 6.2~1012 M-l) is higher than that of the recombinant inhibitor (see Table 2) (Knapp et al., 1992). Moreover, hirudisin interacts directly with glycoprotein IIb-IIIa on platelet surface, suppressing both thrombin- and ADP-induced platelet aggregation (Knapp et al., 1992). This finding demonstrates the importance of the Arg-Gly-Asp motif for the disintegrin activity exerted by inhibitors of blood clotting (e.g. decorsin and ornatin; see Chapter 4) (Krezel et al., 1994).

Biomedical Aspects

The analysis of the anticoagulant effect of hirudin demonstrated that it is a unique highly specific thrombin inhibitor in that it does not affect any related or unrelated (pro)enzyme. When the thrombin-hirudin adduct is formed, all proteolytic activities of the enzyme are blocked. Thus, hirudin prevents not only fibrinogen clotting induced by thrombin, but also further thrombin-catalyzed hemostatic reactions such as activation of clotting Factors V, VIII and XIII, as well as the thrombin-induced platelet reaction. Therefore, by instantaneous inhibition of the small amount of thrombin generated after activation of the coagulation system, the positive feedback on prothrombin activation (that would otherwise lead to accelerated generation of further thrombin) is prevented (Janus et al., 1983; Markwardt, 1985, 1986, 1988; Fenton, 1989; Schnebli and Braun, 1986; Seemtiller et al., 1986; Stone and Maraganore, 1993; Stone, 1995; Stubbs and Bode, 1994).

The last decade has witnessed explosive developments in the pre-clinical and clinical pharmacological applications of hirudin. This has been made possible by the recent application of the recombinant DNA technology, which made several recombinant hirudins available on a large scale for pharmacological screening and clinical trial (Walsmann and Markwardt, 1981; Markwardt et al., 1982; Markwardt, 1956, 1985, 1986, 1988; Harvey et al., 1986; Schnebli and Braun, 1986; Fareed et al., 1989; Walsmann and Kaiser, 1989; Walenga et al., 1990; Marki et al., 1991).

Several manufacturers (in e.g. France, Germany, Italy, Japan, Switzerland, United States and United Kingdom) that provide hirudin have currently the natural or the recombinant inhibitor as well as synthetic hirudin derivatives in Phase I, II or III clinical trials. Therefore, the most potent known natural inhibitor of thrombin is presently gaining popularity as a clinical anticoagulant in the form of native hirudin and of iso-inhibitors including variants with point mutations and N-terminal and C-terminal


modifications. A systematic structural difference with respect to native hirudin prepared from the leech is the absence of the sulfate group from Tyr63 in the recombinant inhibitors (accordingly, they were also named desulfatohirudins) (e.g. Harvey et al., 1986; Marki et al., 1991).

It is the importance of thrombosis in cardiovascular disease that has highlighted the limitations of existing antithrombotic drugs (Badimon et al., 1994). In fact, thrombosis is the predominant cause of early death in industrialized countries (Stubbs and Bode, 1994). Heart attack and stroke are usually caused by the formation of occlusive thrombi in coronary and cerebral arteries at the sites of atherosclerotic stenosis and plaque rupture (Chesebro and Fuster, 1987). Thrombin is the primary mediator of platelet recruitment and activation (i.e. the common event that leads to the formation of localized thrombi) (Wallis, 1988; Kelly et al., 1991; Meyer et al., 1994a). Since the common antithrombotic agents used in therapy (heparin, aspirin and coumarin

Table 5. Pharmacological and clinical aspects of hirudin and hirulog

Pharmacodynamics l Direct inhibition of catalytic and receptor-based thrombin activities l No effect on the biosynthesis of clotting factors l No influence on lipoproteinases l No interaction with platelets and vascular endothelium l No hemorragic side-effects at antithrombotically-effective doses l Very low toxicity after acute or chronic administration l Weak or no immunogenicity

Pharmacokinetics l High bioavailability after subcutaneous administration (plasma level with nearly plateau

values lasting 30 min to 4 hr after administration) l Distribution in the extracellular space (tllZ = 10 min of initial distribution phase after

intravenous injection) 0 No organ deposition 0 Renal excretion in biologically-active form (in the 24 hr urine 50% of the administered

amount) l Half-life of about 1 hr after intravenous injection

Antithrombotic Indications l Unstable angina l Coronary angioplasty (abrupt closure and restenosis) l Deep vein thrombosis after orthopedic surgery l Fibrinolysis l Prevention of reocclusion following thrombolysis in acute myocardial infarction 0 Inhibition of development of localised Schwartzman-Sanarelli phenomenon

Anticoagulant Indications l Hemodialysis and extracorporeal circulation l Disseminated intravascular coagulation l Cardiopulmonary bypass l Antithrombin deficiency


derivatives) are associated with incomplete responsiveness of the disease or with complications (Bar-Shavit et al., 1989; Wilcox et al., 1989; Weitz et al., 1990; Becker, 1992; Cannon et al., 1993; Stone, 1995) hirudin and analogs show promise as novel therapeutic agents (Hirsh, 1991; Badimon et al., 1994; Johnson, 1994; Stone, 1995). In particular, the potential advantages of hirudin and of its small peptidomimetic analogs (such as hirulog) over heparin can be summarized as follows. These macromolecular inhibitors: (1) neutralize thrombin bound to clots or extracellular matrices, which are relatively resistant to heparin (Weitz er al., 1990); (2) do not need a cofactor such as antithrombin III; and (3) are not blocked by activated platelets, which release platelet factor 4 and histidine-rich glycoproteins that neutralize heparin (Badimon et al., 1994; Johnson, 1994).

Investigation of pharmacological properties of hirudin and results of clinical trials are summarized in Table 5 (Markwardt et al., 1982, 1988,1990,1991a, b; Markwardt, 1985; Bichler et al., 1991; Johnson, 1994; Meyer et al., 1994b).

It is relevant that the results of clinico-pharmacological studies in man correspond to those obtained in animal experiments (e.g. dog, rabbit and rat) (Markwardt, 1985). In the coagulation tests, all clotting times are prolonged as a function of the hirudin plasma level, and the overall effect is best followed by thrombin time (Markwardt, 1985). A major concern in the use of these potent thrombin inhibitors is the risk of bleeding, particularly when the antithrombotic therapy is combined with invasive procedures, fibrinolytic treatment, or a patient’s predisposition to abnormal hemostasis. Therefore, to maximize antithrombotic potential while minimizing the risk of excessive bleeding, i.e. in order to have a minimal impact on the systemic coagulation mechanism, the therapeutic action should involve short-term regimens or target the thrombus directly by iocal delivery or selective affinity methods (Johnson, 1994; Meyer et al., 1994b). The question remains of whether a new generation of orally-active thrombin inhibitors can be developed from hirudin to improve the treatment of chronic thrombotic disorders.

Finally, a number of recombinant forms of hirudin are in advanced clinical development. Initial clinical experience indicates that continued infusion (0.5 mgskg-rahr-1 for 30 min) will be necessary in order to achieve a stable degree of anticoagulation (Lid&r et al., 1993).

Chapter 4

Hirustasin

General Aspects

Hirustasin (Hirudo antistasin), an antistasin-like inhibitor of tissue kallikreins, has been prepared recently from leech Hirudo medicinalis (Siillner et al., 1994). Antistasin was originally isolated from the salivary glands of the Mexican leech Haementeria @kinah. A protein almost identical to antistasin has been isolated from the giant Amazonian leech Haementeria ghilianii, and named ghilanten. At variance with hirustasin, antistasin and ghilanten selectively inhibit human Factor Xa activity. Furthermore, a cDNA fragment containing a highly conserved 6-fold repeat with sequence homology to hirustasin, antistasin and ghilanten has been identified in the primitive metaxoan Hydra (Dunwiddie et al., 1993; Hauptmann and Kaiser, 1993a, b; Siillner et al., 1994).

Structural Aspects

Four antistasin isoforms (A, B, C and D) have been isolated from the salivary glands of the Mexican leech Haementeria oficinalis. Antistasin and ghilanten are cysteine-rich polypeptides of 119 amino acids, containing two domains. The amino-acid sequence identity between antistasin isoforms and ghilanten is higher than 90% (see Fig. 2) (Dunwiddie ef al., 1993; Hauptmann and Kaiser, 1993a, b; Sollner et al., 1994).

The amino-acid sequencing of the 55 amino-acid protein revealed that hirustasin is the only antistasin-type protein known to consist of only one domain; 27, 32 and 43% amino-acid identity was found to the antistasin domains 1 and 2, and to the Hydra- antistasin domain 1, respectively. Moreover, the pentapeptide AspGlu-Asn-Gly-Cys (amino-acid residues at positions 40-44 in hirustasin) is completely conserved in hirustasin and in five of six Hydra-antistasin domains (see Fig. 2) (Dunwiddie et al., 1993; Hauptmann and Kaiser, 1993a, b; Krezel et al., 1994; Siillner et al., 1994).

A nearly exact conservation of the spacing of the ten cysteine residues was observed among hirustasin and antistasin, ghilanten and Hydra-antistasin domains (see Fig. 2). Moreover, all these serine proteinase inhibitors display the Cys-c,,-Cys-X-Cys-X -cys-x,-cys-x s_lTCys disulfide bridge pattern observed in hirudin, decorsin, ornatin and epidermal growth factor domains (although in reverse order) (see Fig. 2)

249


(Appella et al., 1988; Dunwiddie et al., 1993; Hauptmann and Kaiser, 1993a, b; Krezel et al., 1994; Siillner et al., 1994).

In hirustasin, the following residues comprise the reactive site loop: Glu26(P,), Va127(P,), His28(P,), Cys29(P,), Arg30(P,), Ile31(P,‘), Arg32(P,‘) and Cys33(P,‘). Residue P, is Gly in the antistasin domain 1, Glu in the ghilanten domain 1, Asp in the antistasin domain 2 and in the ghilanten domain 2, and Lys in the Hydra-antistasin domain 1. Residue P, is Ile in the antistasin domain 2, in the ghilanten domain 2, and in the Hydra-antistasin domain 1. Residue P, is Arg in the antistasin domain 1 and in the ghilanten domain 1, Asn in the antistasin domain 2 and in the ghilanten domain 2, and Gln in the Hydra-antistasin domain 1. Residue P,’ is Val or Met in the isoforms of the antistasin domain 1, Val in the ghilanten domain 1, Lys in the antistasin domain 2 and in the ghilanten domain 2, and Met in the Hydra-antistasin domain 1. Residue P2’ is His in the antistasin domain 1, Tyr in the ghilanten domain 1, Thr in the antistasin domain 2 and in the ghilanten domain 2, and Phe in the Hydra-antistasin domain 1 (see Fig. 2) (Dunwiddie et al., 1993; Krezel et al., 1994; Sollner et al., 1994).

In hirustasin, the P, reactive site residue has been identified as Arg30, corresponding to Arg34 in the inhibitory N-terminal antistasin and ghilanten domain 1 (see Fig. 2), being consistent with trypsin-like serine proteinase inhibitory properties (see Tables 6 and 7) (Hofman et al., 1992; Dunwiddie et al., 1993; Hauptmann and Kaiser, 1993a, b; Sollner et al., 1994).

The Lys90 residue present in the C-terminal domain of antistasin and ghilanten has been postulated to play a significant role in stabilizing the interaction between the

Table 6. Values of thermodynamic parameters for hirustasin and antistasin binding to serine proteinases

K (M-l)

Proteinase Hirustasin Antistasin

Bovine P-trypsin* 1.4 x 10s 2.0 x loq Bovine a-chymotrypsin* 1.6 x 108 <2x105 Human cathepsin G* 3.4 x 10s n.d. Porcine tissue kallikrein* 7.7 x 107 nd.

Human plasmin* 7.2 x 106 nd. Human urokinase* <1x105 nd. Human plasma kallikrein* <lx105 n.d. Human thrombin* <lx105 < 2 x 105

Human Factor Xa < 1 x 105* 2.0 x 1oq

Porcine pancreatic elastase* < 1 x 10s < 2 x 10s

Human leukocyte elastase* <lx105 <2x105

Bacillus subtilis subtilisin * <lx105 n.d. Porcine chymase* c 1 x 105 n.d.

*pH 7.8 and 25°C from Sdllner et al.. (1994); tICso value; $pH 8.2 and 22°C from Hofmann et al. (1992); n.d., not determined.

Proteinase Inhibitors from Leech Mudo medicinalis 251

Table 7. Values of thermodynamic and kinetic parameters for wild-type and mutant antistasin binding to human Factor Xa (pH 8.2 and 22”C)*

Antistasin K (M-l) k,, (M-l secl) k,ff (secl)

Wild-type 2.0 x 1010 1.2 x 107 6.4 x 10-4 Arg34 + Lys 7.8 x 108 2.9 x 106 4.2 x lo-3 Va13.5 + Ile 5.6 x 109 2.0 x 107 4.0 x lo-3 Lys90 + Val 1.0 x 10’0 7.4 x 106 7.5 x lo-4 Lys90 -+ Ile 1.2 x 10’0 7.9 x 106 6.8 x 1v

*From Hofmann ef al. (1992).

serine proteinase inhibitor and Factor Xa, when Factor Xa is assembled in the prothrombinase complex. Moreover, a C-terminal heparin and sulphatide binding domain (Pro93-Ser119) has been identified in antistasin and ghilanten (Dunwiddie et al., 1993; Hauptmann and Kaiser, 1993a, b; Sollner et al., 1994). Crystals of antistasin have been reported (Schreuder et al., 1993), and the determination of its structure is well under way (R. Lapatto, B. W. Dijkstra and W. G. J. Hol, personal communication; Stubbs and Bode, 1994). It has been proposed that antistasin exhibits a fold similar to that of hirudin and decorsin, a potent inhibitor of platelet aggregation from leech Mucrobdelh decor-u (Krezel et al., 1994).

Detailed studies, including cDNA cloning are necessary to clarify whether hirustasin is a fragment of a multi-domain precursor protein which may have evolved by gene duplication events similarly to antistasin (Sdllner et al., 1994).

Functional Aspects

Hirustasin, antistasin and ghilanten obey the ‘standard mechanism’ of proteinase inhibition (see Scheme 2) both from the thermodynamic and kinetic viewpoints (Hofmann et al., 1992; Dunwiddie et al., 1993; Hauptmann and Kaiser, 1993a, b; Sollner et al., 1994).

Hirustasin is the first inhibitor of tissue kallikrein identified in leeches, and is also a tight-binding inhibitor of bovine l3-trypsin, bovine cll-chymotrypsin, human cathepsin G and human plasmin (see Table 6). However, despite the high similarity to antistasin and ghilanten particularly in the vicinity of the reactive site bond (Arg30-Ile31), hirustasin does not affect human Factor Xa action (see Table 6). In this respect, hirustasin neither affects the prothrombin coagulation time nor the partial thromboplastin coagulation time, in vitro. Thus, factors other than the reactive-site sequence (i.e. the Ps-Ps’ region) appear to contribute significantly to the specificity of antistasin-like serine proteinase inhibitors (see Table 6) (Siillner et al., 1994).

Wild-type and recombinant antistasin display the same inhibitory effect towards human Factor Xa and related serine proteinases (see Tables 6 and 7) (Hofmann et al., 1992; Siillner et al., 1994).


Site-directed mutagenesis studies on antistasin demostrate the requirement for a positively-charged residue at position Pi, with Arg preferred to Lys (see Table 7). Mutations at positions P, (Arg34 -+ Asn, and Arg34 -+ Leu) and Pi’ (Vat35 -+ Ile) reduced and abolished the inhibitory potency of antistasin towards human Factor Xa (see Table 7). Moreover, mutation at position 90 (Lys90 + Val, and Lys90 + Ile), is expected to mimic the reactive site present in the antistasin and ghilanten domain 1 or the cleavage site in prothrombin (see Fig. 2), does not affect significantly human Factor Xa inhibition (see Table 7). This behaviour indicates that, although the N- and C-terminal domains of antistasin share significant amino-acid sequence similarity (see Fig. 2), the antistasin domain 2 is not folded in a conformation which facilitates strong binding with human Factor Xa (Hofmann et al., 1992; Dunwiddie et al., 1993; Hauptmann and Kaiser, 1993a, b; Sollner et al., 1994).

The clotting inhibitory profiles of the antistasin mutants spanning the Pi to Pi’ positions mirror the results given in Table 7. In contrast, the two antistasin mutants at position 90 (Lys90 -_) Val, and Lys90 + Ile) exhibit considerably reduced potencies approaching that of Arg34 + Lys in clotting inhibition, whereas they are only marginally less potent than wild-type antistasin in the chromogenic human Factor Xa assay. These results indicate that residue 90 may play a significant role in stabilizing the interaction of antistasin with Factor Xa when this serine proteinase is assembled into the multimeric prothrombinase complex (Hofmann et al., 1992).

Values of the second-order rate constant for the human Factor Xa-wild-type and Factor Xa-;mutant antistasin complex formation (i.e. k,, values) range between 2.9~106 M-%ecF1 and 2.0~10~ M-l set-1; on the other hand, values of the first-order rate constant for the enzyme-inhibitor complex dissociation range between 6.4x10-4 and 4.2x10-3 set-1 (see Table 7) (Hofmann et al., 1992). According to Ascenzi et al. (1992a) and Cutruzzola’ et al. (1993), this behaviour indicates that the accessibility of antistasin to human Factor Xa (expressed by k,, values) may depend on the mutant considered. Moreover, kinetics of the human Factor Xa- antistasin complex dissociation (qxpressed by k,, values), indicate that domain(s) protruding into the human Factor Xa active site area and strongly contacting the inhibitor should contribute to the binary complex destabilization (Ascenzi et al., 1992a).

The formation of the stable inactive binary complexes of human Factor Xa with native and recombinant antistasin involves two steps: a fast binding reaction followed by a rate-limiting slow isomerization process (Hofmann et al., 1992).

Biomedical Aspects

Hirustasin selectively inhibits tissue kallikreins. On the other hand, antistasin and ghilanten are potent specific inhibitors of the blood coagulation Factor Xa. In this respect, selective Factor Xa inhibition by recombinant antistasin: (1) prevents vascular graft thrombosis in baboons and rabbits; (2) accelerates the reperfusion and prevents reocclusion in a canine model of femoral arterial thrombosis; (3) reduces restenosis after balloon angioplasty of atherosclerotic femoral arteries in rabbits; and (4) affects the mitosis of cultured aortic smooth muscle cells (Dunwiddie et al., 1993; Hauptmann and Kaiser, 1993a, b; Sdllner et al., 1994).

Proteinase Inhibitors from Leech Hirudo medicinelis 253

Antistasin also displays marked antimetastatic properties. In this respect, it may be recalled that the metastatic spread of tumours is correlated to abnormal stimulation of blood coagulation, possibly due to direct activation of Factor X by tumour cells. Moreover, since the deposition of fibrin provides the tumour cells with a protective covering, impermeable to the immune system, the effect of antistasin on tumour spread might be exerted via Factor Xa inhibition (Dunwiddie et al., 1993; Hauptmann and Kaiser, 1993a, b; Sollner et al., 1994).

The antithrombotic efficacy and duration of action of a single subcutaneous administration of recombinant antistasin was evaluated in a Rhesus monkey model of mild disseminated intravascular coagulation. Antistasin (1 mg/kg) was shown to be fully effective and comparable to standard heparin (1000 U/kg) in the suppression of the thromboplastin-induced fibrinopeptide A generation for at least 5 hr following a single subcutaneous administration. The absorption rate of antistasin mirrored that of standard heparin exhibiting peak anticoagulant activity between 1 and 2 hr post-administration. The anticoagulant effects of a single antistasin dose lasted for longer than 30 hr. Since repeated subcutaneous administration of antistasin resulted in fully neutralizing antibodies, the Factor Xa inhibitor is probably not suitable for chronic subcutaneous anticoagulant therapy (Dunwiddie et at., 1993; Hauptmann and Kaiser, 1993a, b).

Eglin

General aspects

Eglins (Elastase-Cathepsin G Leech Inhibitors) are small proteins present in the medicinal leech Hirudo medicinalis, with strong inhibitory activity against chymotrypsin- and subtilisin-like serine proteinases acting on non-cationic substrates (Seemiiller et al., 1986; Ascenzi et al., 1991).

Eglins are distributed over the whole body of the leech (one leech contains approximately 20 p.g of eglin), and particularly in the blood lacunae, testicles, mucus, mucus cells and nephridia (Seemiiller et al., 1986). Although, the biological function of eglins in the leech has not been completely clarified, a hypothetical role of the inhibitors in preservation of the blood stored in the foregut has been proposed (Roters and Zebe, 1992).

Because of their specificity, efficiency or potency, low toxicity and ‘bioavailability’, eglins have a therapeutic potential in the pathogenesis of pulmonary diseases and inflammatory processes involving leukocyte elastase and cathepsin G, considered as the main target enzymes. Cross-inhibition of other serine (pro)enzymes by eglins may probably be of little pharmaceutical consequence (Schnebli and Braun, 1986; Seemtiller et al., 1986; Ascenzi et al., 1991). Eglin-like serine proteinase inhibitors have been isolated not only from the leech Hirudo medicinalis, a species which feeds on frogs and mammals (Seemtiller et al., 1977, 1986) but also from the leech Hirudinaria manillensis, a species specialized in feeding on mammals, especially water buffalos (Electricwala et al., 1993b), and from two non-blood sucking species of North American leeches Herpobdella punctata and Nephelopsis obscura (Goldstein et al., 1986). Leukocyte elastase, cathepsin G and proteinase-3 inhibitors from different species of leech, Hirudo, and Hirudinaria are functionally similar, but are fundamentally different in their physico-chemical properties (Electricwala et al., 1993b). Thus, eglin-like serine proteinase inhibitors belonging to different leech species may not have evolved from a common ancestral gene (Electricwala et al., 1993b).

255


Structural Aspects

The eglins may be prepared from the whole leech (Seemtiller et al., 1977, 1986), and from leech dilute saliva (Rigbi et al., 1987a). Eglin c has been obtained in Escherichiu coli (Rink et al., 1984), and prepared by organic synthesis (Okada et al., 1990a).

Eglin exists in multiple forms. The two main components, eglin b and eglin c, showing nearly identical M, values of 8073 and 8099 Da, respectively, are eluted together in fraction IIb,c from the ion-exchange chromatography DEAE-cellulose column; two other forms (in fraction IIe and IIf) are eluted by the salt gradient. Very close isoelectric points have been observed for the two main inhibitor components (6.6 and 6.45 for eglin b and c, respectively) (Seemtiller et al., 1977, 1986). Eglin b and c are single-chain polypeptides, each composed of 70 amino-acid residues. The two inhibitors show an identical pattern of residue amidation. Position 35 is occupied by histidine in eglin b instead of tyrosine in eglin c (see Fig. 6) (Seemtiller et al., 1986). The recombinant eglin c differs from the native wild-type inhibitor at its N-terminus, the Thrl residue being modified by acetylation of the cc-amino-group (Rink et al., 1984).

Eglin b and c are members of the potato inhibitor I family of serine proteinase inhibitors (Laskowski and Kato, 1980) that also includes yeast proteinase B inhibitor (16% amino-acid identity) (Maier et al., 1979), barley chymotrypsin inhibitors CI-1 and CI-2 (22 and 26% amino-acid identity, respectively) (Jonassen and Svendsen, 1982), and the potato inhibitor I (36% amino-acid identity) (Richardson and Cossins, 1974) (see Fig. 6). Generally, serine proteinase inhibitors belonging to the potato inhibitor I family are characterized by the absence of disulfide bridges and cysteine residues in their structures (see Fig. 6), a particularly unusual property for protein proteinase inhibitors obeying the ‘standard mechanism’ of inhibition (see Scheme 2) (Laskowski and Kato, 1980; Read and James, 1986; Bode and Huber, 1991, 1992). In this respect, eglin b and c have been designated ‘cysteine-independent inhibitors’ by M. Laskowski, Jr (Seemtiller et al., 1986). The only exception in the family is represented by the potato inhibitor I which displays one disulfide bridge (see Fig. 6) (Richardson and Cossins, 1974). However, potato inhibitor I properties are not affected by reduction and carboxymethylation of cystine residues (Seemiiller et al., 1986).

Despite the absence of intramolecular covalent cross-bridges, eglins are very stable molecules which retain their full biological activity between pH 2 and 9 at 80°C for 10 min. Degradation at the C-terminus by carboxypeptidase Y only occurs in the presence of 0.5% sodium dodecyl sulphate. Digestion with trypsin is possible only after acid denaturation with 20% trifluoroacetic acid at 70°C for 1 hr (Seemtiller et al., 1986).

The three-dimensional structures of free native eglin c, of eglin c with hydrolyzed reactive centre, of the Leu45 + Arg, Asp46 + Ser eglin c double mutant, and of the native eglin-homologous barley inhibitor CI-2 have been solved (Clore et al., 1987a, b; Hyberts et al., 1987, 1992; Heinz et al., 1992; Hipler ef al., 1992; Hyberts and Wagner, 1990; Peng and Wagner, 1992; Wagner et al., 1992; Betzel et al., 1993). The crystal structures of the binary complexes of eglin c with Bacillus licheniformis subtilisin (B. licheniformis subtilisin; subtilisin Carlsberg), with Bacillus subtilis subtilisin (B. subtilis subtilisin; subtilisin Novo), with the thermostable alkaline


YPBI BI CI-1 BI CI-2

Eglin c YPBI BI CI-1 BI CI-2 PI I

Bglin c YPBI BI CI-1 BI CI-2 PI I





Thr.Glu~Phe.Gly~Ser~Glu.Leu Lys ---.Ser*Phe.Pro.Glu* Lys~Aen~Phe~Ile+Val~Thr~Leu Lys Lys*Asn.Thr.Pro.Asp.

l-l Ala.Ser.Gly.---.------Ala LYS Thr.Ser*Tm.Phe.Glu. Ala.Gly.Asp.Arg*HisAsnLeu Thr.Glu.TG.Pro.Glu. Lye.Glu.Phe.Glu.Cys+Asp.Gly Leu.Gln.Trp.Pro.Glu.

Fig. 6. Amino-acid sequences of eglin c, of yeast proteinase B inhibitor (YPBI), of barley inhibitors CI-1 and CI-2 (BI CI-1, type C, and BI CI-2, type A, respectively) and of the potato inhibitor I (PI I, type A) (Richardson and Cossins, i974; Maier et al., 1979; Jonassen and Svendsen, 1982; McPhalen et al., 1985b; Seemtiller et al., 1986). The arrow marks the Pt-P,’ reactive site bond. Gaps are indicated by dashes. Identical residues are boxed. Multiple alignments were produced using the Genetics Computer Group sequence analysis software

package (GCG version 7.1) using a VaxNMS system (Devereux et al., 1984).

protease from Thermoactinomyces vulgar-is (thermitase), with mesentericopeptidase and with bovine a-chymotrypsin, of the Leu45 -+ Arg and Arg53 + Lys eglin c mutants with B. subtih subtilisin, and of the eglin-homologous barley inhibitor CI-2 with B. sub& subtilisin have also been determined (McPhalen et al., 1985a, b; Bode et al., 1986, 1987; McPhalen and James, 1987, 1988; Dauter et al., 1988, 1991; Gros et al., 1989a, b, 1992; Bolognesi et al., 1990; Heinz et al., 1991, 1992; Frigerio et al., 1992). Finally, the binary complex of human leukocyte elastase with the Thr44 + Pro eglin c mutant has been crystallized, but the three-dimensional structure has not been reported (Heinz et al., 1989).

P. Ascenzi et al.

.CD1 145 .CDl 145

148 148

Fig. 7. Stereo view of the Ca backbone of eglin c as observed in the binary bovine a-chymotrypsin-inhibitor complex (Frigerio et al., 1992). The Leu45(P,, for this symbol see Fig. 1) residue, conferring the serine (pro)enzyme primary specificity to eglin c, is shown in the upper part of the picture. The following residues are involved in the stabilization of the inhibitor reactive site loop: Thr44(P,), Asp46(P,‘), Arg48(P,‘), ArgSl, Arg53 and the Gly70 C-terminal carboxylate. ArgSl and Arg53 are contributed by the central P-sheet region. The

inhibitor residues are identified by the suffix ‘I’.

Proteinase inhibitors belonging to the potato inhibitor I family, including eglin c, show a compact, wedge-shaped three-dimensional organization of the polypeptide chain (see Fig. 7). The secondary structure of eglin c consists of a short 3,, helical tract (residues PhelO-Val14) followed by 3.4 turns of regular a-helix (residues ValWTyr29), and of a central four-strand mixed parallel/antiparallel P-sheet (the four strands covering residues Lys8-PhelO, Asn33-Leu37, Arg51-Tyr56 and His66Gly70). The a-helix and the P-sheet, together with tight p-turns, define a rather compact and apolar core region of the inhibitor molecule which supports the reactive site, an extended polypeptide loop comprising residues Pro38Asn50. The Thrl-Leu7 peptide has been observed only in the free native eglin c crystal structure where it adopts an extended conformation protruding from the molecular surface (Hipler et al., 1992). Alternatively, the N-terminal heptapeptide has not been observed in the three-dimensional structures of free eglin c with hydrolysed reactive centre, and in the serine proteinase-inhibitor complexes, owing to proteolytic digestion by the target enzyme during crystallization or due to conformational disorder (Bode et al., 1986, 1987; Bolognesi et al., 1990; Dauter et al., 1991; Betzel et al., 1993). In this respect, nuclear magnetic resonance experiments indicate a high mobility of the native eglin c N-terminal heptapeptide in solution (Hyberts et al., 1992).

In eglin b and c, the following residues comprise the reactive site loop: Gly40(P,), Ser41(P,), Pro42(P,), Va143(P,), Thr44(P,), Leu45(P,), Asp46(P,‘), Leu47(P,‘), Arg48(P,‘) and Tyr49(P,‘). Also residues Phe55(P,,‘) and His68(P,,‘) participate in the serine (pro)enzyme-eglin c recognition processes (Ascenzi et al., 1991; Frigerio et al., 1992). Residues forming the neighbourhood of the reactive site (P3-P3’) adopt a polypeptide backbone conformation comparable with that observed in the same region of other protein serine proteinase inhibitors (Read and James, 1986; Bode and Huber,

Proteinase Inhibitors from Leech Himdo medicinalis 259

1991, 1992; Stubbs and Bode, 1994), consistent with the expected conformation for an ideal serine proteinase substrate (Hipler er al., 1992). Leu45(P,) is roughly at the centre of the proteinase binding loop and points towards the incoming cognate enzyme (i.e. towards the solvent in the free inhibitor) (see Fig. 7) (Hipler et al., 1992). Leu45(P,) is the primary determinant of the inhibitor specificity, consistent with eglin’s chymotrypsin- and subtilisin-like enzyme inhibition pattern (see Tables 89, 10 and 11) (Fink et al., 1986a; Seemtiller et al., 1977, 1986; Braun et al., 1987; Faller et al., 1990; Ascenzi et al., 1991; Junger et al., 1992; Heinz et al., 1992; Cutruzzola’ et al., 1993).

The eglin c reactive site loop is devoid of covalent cross-links to the inhibitor molecular core, being stabilized in its three-dimensional structure by strong polar interactions. In particular, residues Thr44(P,), Asp46(P,‘) and Arg48(P,‘) make part of an interlaced hydrogen-bonded network connecting the reactive site loop to residues of the central p-sheet (ArgSl and Arg53) and to the Gly70 C-terminal carboxylate (see Fig. 7). A total of 16 intramolecular hydrogen bonds are observed in this inhibitor region (including two tightly bound water molecules) (see Fig. 7) (Bode et al., 1987; McPhalen and James, 1988; Gros et al., 1989b; Frigerio et al., 1992; Hipler et al., 1992).

Despite these clearly defined polar interactions, the reactive site loop of eglin c demonstrates a perceivable degree of structural flexibility (Dauter et al., 1991; Frigerio et al., 1992; Hipler er al., 1992; Hyberts et af., 1992). Two processes, hinge bending

Table 8. Values of thermodynamic parameters for the binding of native eglin b and c, recombinant eglin c (r-eglin c) and synthetic eglin-c (s-eglin c) to serine proteinases

Proteinase

Human cathepsin G

Human leukocyte elastase

Bovine wchymotrypsin

B. licheniformis subtilisin

Inhibitor

native eglin b* native eglin c* r-eglin cf s-eglin c$ native eglin b* native eglin c* r-eglin CT s-eglin c$ native eglin b* native eglin c* r-eglin co s-eglin cl1 native eglin b* native eglin c* r-eglin CT

K(M-1)

4.0 x 109 3.6 x 109 7.0 x 109 2.5 x 109 4.3 x 109 5.0 x 109 1.0 x 10’0 1.2 x 10’0 3.3 x 109 1.4 x 109 2.1 x 109 2.4 x 109 5.0 x 109 8.3 x 109 6.6 x 109

*pH 7.8 and 25”C, from Seemtiller et al. (1977); tpH 8.0 and 21°C from Ascenzi et al. (1991); +pH 8.0 and 21°C values of K (from Okada et al., 1990a) have been normalized to those reported by Ascenzi et al. (1991); §pH 7.0 and SC, from Cutruzzola’ et al. (1993); [IpH 7.0 and SC, values of K (from Okada et al. (1990a) have been normalized to those reported by Cutruzzola’ et al.

(1993).


Table 9. Values of thermodynamic and kinetic parameters for eglin c binding to serine (pro)enzymes

Proteinase k ($1 sect)

k off (se+)

Leu-proteinase’ Human cathepsin G Rat cathepsin GS Human leukocyte elastase Human leukocyte elastases

(with heparin) Human leukocyte elastase()

(rqmacroglobulin-bound) Baboon leukocyte elastaseq Ovine leukocyte elastasen Rat leukocyte elastaseg Human leukocyte proteinase-3+ B. subtilis subtilisin# B. licheniformis subtilisin* ThermitaseV Bovine a-chymotrypsino Bovine Met(O)192 cu-chymotrypsino Bovine or-chymotrypsinogen A* Bovine P-trypsin* Porcine chymaseo Human pancreatic elastaseo Porcine pancreatic elastase*

2.2 x 10” 7.0 x 109* > 1 x 106 1.0 x lolo* n.d.

1.9 x 106

n.d. nd. >lx108 nd. 3.1 x 1010 6.6 x 109 2.0 x 10’0 2.1 x 109 3.5 x 108 6.5 x 105 1.2 x 105 2.3 x 10’ 2.7 x 109 3.6 x 10s

n.d. 2.0 x 106t nd. 1.0 x lo’? 4.7 x 104

1.1 x 103

2.5 x 107 2.2 x 105 nd. 4.2 x 104 n.d. n.d. n.d. 2.4 x 105 1.1 x 105 nd. n.d. 2.5 x 106 7.3 x 105 n.d.

n.d. n.d. n.d. 1.0 x lO-st n.d.

5.9 x lo”

n.d. n.d. n.d. n.d. n.d. n.d. n.d. 1.2 x 10-4 3.1 x lo” n.d. n.d. 1.1 x lo-1 2.7 x lO-’ n.d.

*pH 8.0 and 21”C, from Ascenzi et al. (1991); tpH 7.3 and 37°C from Braun et al. (1987); $pH 7.5 and 37”C, from Virca et al. (1984); $pH 7.5 and 25°C from Frommherz et al. (1991); JlpH 7.5 and 25°C from Stack et al. (1988); (pH 7.4 and 37°C from Junger et al. (1992), +pH 7.5 and 25% from Rao et al. (1991); #pH 7.8 and 37°C from Heinz et al. (1992), VpH 8.0 and 25”C, from Krystek et al. (1993); opH 7.0 and 5°C from Cutruzzola’ et al. (1993); OpH 7.8 and 25”C, from

Fink et al. (1986a); OpH 7.4 and 25”C, from Faller et al. (1990); n.d.: not determined.

around Pro42(P,) and Tyr49(P,‘) sites and conformational changes in the neighbourhood of the Leu45(P,)-Asp46(P,‘) scissile peptide bond, are used in order to attain proper enzyme inhibitor contacts and stabilize the intermolecular interactions. The degree of conformational readjustement achieved in the reactive site loop varies in the different serine proteinase-eglin c complexes analyzed (2-4 A; Ca displacement at the reactive site loop) (see Fig. 8) (McPhalen and James, 1988; Frigerio et al., 1992). A hinge bending motion (3 A) of this region of eglin c has been observed also in solution (Hyberts et

al. , 1992).

The flexibility of the inhibitor reactive site can be related to the absence of covalent intramolecular cross-links (i.e. disulfide bridges) supporting the primary (pro)enzyme contact loop, which are observed in almost all serine proteinase protein inhibitor families (Laskowski and Kato, 1980; Read and James, 1986; Bode and Huber, 1991, 1992; Stubbs and Bode, 1994), and may reflect the broad inhibitor specificity of eglin

Proteinase Inhibitors from Leech Hirudo medicinaiis 261

Table 10. Values of thermodynamic parameters for wild-type and mutant eglin c binding to serine proteinases

Proteinase Eglin K (M-l)

Human leukocyte elastase

Bovine P-trypsin

B. subtilis subtilisint

wild-type* Thr44 --+ Prot Arg.51 -+ Lyst Arg53 + Lyst wild-type* Leu45 + Arg$ wild-type Leu45 + Arg Arg53 --+ Lys

1.0 x 10’0 7.2 x 108 2.8 x 107 1.8 x 10s 1.2 x 105 4.0 x 10’0 3.1 x 10’0 1.9 x 10’0 9.1 x 109

*pH 8.0 and 21”C, from Ascenzi et al. (1991); fpH 8.0 and 21”C, values of K (from Heinz et al.,

1992) have been normalized to those reported by Ascenzi et al. (1991); $pH 7.8 and 37”C, from Heinz et al. (1992).

Table 11. Values of thermodynamic parameters for the binding of eglin c and of the inhibitor peptide fragments to serine proteinases

K (M-l)

Inhibitor Human

Leukocyte Elastase Human

Cathepsin G Bovine

a-Chymotrypsin

Eglin c Lys&Gly70 Arg22-Glyn70 Gln3 l-Gly70 Ser41-Gly70 Leu45-Gly70 AsnSO-Gly70 Thr60-Gly70 Arg22-Phe25 Gln31-Tyr35 Gln31-Gly40 Ser41-Tyr49 Thr60-Va163

1.0 x 1orc* 2.7 x lO’c$ 2.1 x 107* 3.3 x 107$ 4.2 x 105$ 8.0 x 1oq 2.7 x lOs$ 4.2 x 104:: n.d. 9.6 x 1Oq n.d. n.d. 1.2 x 1041)

7.0 x 109* 2.1 x lost 5.0 x 1oq 5.2 x 109Q 5.6 x lay 3.3 x 1059 1.2 x 107$ 6.6 x 1050 3.5 x 1053: 1.1 x lW§ 1.0 x 102$ 3.3 x 1050 1.0 x 1oq 1.7 x 105§ 1.0 x 1oq 7.98 2.3 x 10311 1.4 x 1Oq 2.7 x 10311 n.d. 8.4 x 10311 nd. 4.9 x 10411 5.0 x 1q n.d. n.d.

*pH 8.0 and 21”C, from Ascenzi et al. (1991); tpH 7.0 and 5”C, from Cutruzzola’ et al. (1993); $pH 8.0 and 21”C, values of K (from Okada et al., 1990b) have been normalized to those reported by Ascenzi et al. (1991); 8pH 7.0 and 5”C, values of K (from Okada et al., 1990b) have been normalized to those reported by Cutruzzola’ et al. (1993); JlpH 8.0 and 21”C, values of K (from Okada et al., 1989) have been normalized to those reported by Ascenzi et al. (1991); VpH 8.0

and 25”C, from Okada et al. (1989); n.d.: not determined.


Fig. 8. Stereo view of the best-fit overlay of the Co backbone structures of eglin c, as observed in the binary bovine a-chymotrypsin- and B. licheniformis subtilisin-inhibitor complexes (continuous and dashed lines, respectively) (Bode et al., 1986, 1987; Frigerio et al., 1992). The Ca atom of the Leu45(P,) residue is labelled in the upper part of the picture. The inhibitor residue is identified

by the suffix ‘I’.

c, displaying almost the same affinity (i.e. values of K) for different serine proteinases (see Table 9) (Seemtiller et al., 1986; Ascenzi et al., 1991).

Absence of covalent intramolecular cross-bridges is reflected in the substantial structural readjustments observed in the nicked eglin c, whose crystal structure shows opening of the reactive site with residues Gly40-Leu45 projecting away from the main body of the molecule in an extended conformation (Betzel er al., 1993). A less pronounced effect has been observed in cleaved serine proteinase ovomucoid inhibitors which display only some disorder of the residues immediately surrounding the cleaved reactive site, due to the presence of intramolecular disulfide bridges supporting the (pro)enzyme binding loop (Musil et al., 1991).

Serine proteinase-eglin c complexes display the classical mushroom-shaped overall organization, the inhibitor representing the stalk and the enzyme the head (see Fig. 9) (Huber and Bode, 1978; McPhalen et al., 1985a; Read and James, 1986; Bode et al., 1986, 1987; McPhalen and James, 1988; Gros et al., 1989a, b, 1992; Bode and Huber, 1991, 1992; Heinz et al., 1991, 1992; Frigerio et al., 1992; Stubbs and Bode, 1994).

Inspection of the contact region in the serine proteinase-eglin c complex shows that the inhibitor reactive site loop (the thin rim of the wedge-shaped molecule) is deeply buried in the proteinase active centre (a cleft defined by the two (pro)enzyme domains) (see Fig. 9). As a result, about 1500 AZ of molecular surface becomes solvent inaccessible and approximately 130 intermolecular van der Waals contacts (c4.0 A) are established between eglin c and serine proteinases belonging to the chymotrypsin and subtilisin superfamilies (Bode et al., 1986, 1987; Bolognesi et al., 1987, 1990; McPhalen and James, 1988; Frigerio et al., 1992). These figures are in good agreement with the statistics on the structure of protein-protein recognition complexes (Janin and Chothia, 1990).

Proteinase inhibitors from Leech Hirudo medicinalis 263

Panel A

Panel B

Fig. 9. Stereo view of the Ca backbone of the bovine u-chymotrypsin-eglin c and thermitase-eglin c complexes (panels A and B, respectively) in two orthogonal orientations (Frigerio et al., 1992; Gros et al., 1992). The inhibitor is in the lower part of the picture and highlighted in thick

bonds.

Analysis of residues that control recognition of proteinase protein inhibitors in the chymotrypsin superfamily has demonstrated that, in general, these residues are distributed over eight polypeptide segments on the (pro)enzyme surface (Creighton and Darby, 1989). Six of these (comprised between residues 35-43,5&65,140-153,168-178, 188-196 and 212-221) are contacting eglin c in its complex with bovine a-chymotrypsin (Frigerio et al., 1992). The enzyme-inhibitor contact is asymmetric, and more extensive interaction is observed on one of the two faces of the wedge-shaped inhibitor (see Fig 9, panel A) (Frigerio et al., 1992).

264 P. Ascenzi etal.

In the binary complexes of eglin c with serine proteinases belonging to the subtilisin superfamily contacts between the inhibitor and the target enzyme are limited almost exclusively to the reactive site loop residues (P,-P,‘). In the enzyme- inhibitor complex, the eglin c segment Pro42-Leu45 (P4P1) is the central strand in a small three-stranded anti-parallel P-sheet which includes the proteinase residues 100-102 and 125-127. As a general property of eglin c complexes with subtilisin-like serine proteinases, different orientations of the inhibitor molecule with respect to the target enzyme (up to 100) have been observed, reflecting eglin c flexibility at the reactive site loop (see Figs 8 and 9, panel B) (Bode et al., 1986, 1987; McPhalen and James, 1988; Gros et al., 1989b, 1992; Dauter et al., 1991; Heinz et al., 1991).

Besides the several van der Waals contacts that take place throughout the enzyme-inhibitor contact interface, and which contribute to the enthalpy of the binding process, specific and directional interactions occur mainly at the bovine cr-chymotrypsin and eglin c (re)active site region, involving the P, to P,’ and P,, ’ inhibitor subsites (Frigerio et al., 1992). Protein-protein contacts occur first at the primary recognition centre, involving the proteinase S, and eglin c P, subsites, where the apolar primary specificity pocket of bovine a-chymotrypsin (S,) hosts the side-chain of the Leu45(P,) eglin c residue (see Fig. 10) (Frigerio et al., 1992). As expected from the strong polypeptide backbone structural conservation, the bovine or-chymotrypsin- eglin c hydrogen-bonding pattern, occurring in the surroundings of the primary specificity subsite, is in excellent agreement with those reported for other serine (pro)enzyme-inhibitor complexes (Read and James, 1986; Bolognesi et al., 1987; Bode et al., 1991, 1992; Frigerio et al., 1992; Stubbs and Bode, 1994), and is identical with that observed in the bovine rw-chymotrypsin-turkey ovomucoid inhibitor complexes, between the P, and P,’ inhibitor subsites (Fujinaga et al., 1987; Frigerio et al., 1992). These interactions essentially occur between backbone hydrogen-bond donors or acceptors (of the two partner proteins), while further away from the inhibitor reactive site (beyond sites P, and P3’), the eglin c recognition loop contacts bovine a-chymotrypsin, and stabilizes the complex through interactions based on side-chains contacts, part of which can also be mediated by bridging water molecules. In this respect, a hydrogen bond is additionally observed between the eglin c His68(P,,‘) residue and Tyr146 present in the autolysis loop of bovine cl-chymotrypsin (Frigerio et al., 1992).

Fig. 10. Stereo view of the reactive site loop of eglin c (in the lower part of the picture) and of the bovine a-chymotrypsin active site region (Frigerio et al., 1992). The eglin c Leu45(P,) residue and the serine proteinase Ser195 side-chain are in the central part of the picture. The

enzyme and inhibitor residues are identified by the suffix ‘E’ and ‘I’, respectively.

Proteinase inhibitors from Leech Hitudo medicinalis 265

Ordered water molecules at the enzyme-inhibitor interface have been observed regularly in the complexes studied by X-ray crystallography. In the case of eglin c adducts with serine proteinases belonging to the subtilisin superfamily, due to the poor structural complementarity between the partner proteins, the contact interface is particularly rich in such ordered solvent sites, which participate to the molecular recognition process even at sites immediately surrounding the reactive site (Bode et al., 1987; McPhalen and James, 1988).

The binding of the eglin c Leu45(P,)-Asp46(P,‘) scissile petide bond to the active centre of bovine a-chymotrypsin (taken as the prototype enzyme of the homonymous serine proteinase superfamily) follows the stereochemical rules observed for ‘small’ low molecular weight serine (pro)enzyme inhibitors obeying the ‘standard inhibition mechanism’ (see Scheme 2) (Laskowski and Kato, 1980; Bode and Huber, 1992; Frigerio et al., 1992) (see Fig. 10). Tight packing of residues is observed around the catalytic triad of inhibited bovine cr-chymotrypsin (see Fig. 10) (Frigerio et al., 1992). As observed for the other protein serine proteinase inhibitors (Read and James, 1986; Bode and Huber, 1991, 1992; Stubbs and Bode, 1994), the eglin c reactive site Leu45(P,) carbonyl oxygen is hydrogen bonded to peptide N-atoms from the bovine a-chymotrypsin Gly193 and Ser195 residues (forming the so called ‘oxyanion binding hole’) (see Fig. 10) (Frigerio et al., 1992). The carbonyl atom of the potentially scissile Leu45(P,)-Asp46(P,‘) peptide bond is in tight van der Waals contact (2.75 A) with the proteinase nucleophile Ser195 OG atom , consistent with the proteolytic cleavage of the inhibitor reactive site bond (led by the enzyme Ser195 OG atom) frozen at an incipient stage of the reaction (see Fig. 10) (Huber and Bode, 1978; Bolognesi et al., 1982; Marquart et al., 1983; Read and James, 1986; Bode and Huber, 1991, 1992; Frigerio er al., 1992; Stubbs and Bode, 1994). Very similar considerations have been developed in parallel for the (re)active site interactions occurring in the binary complexes of eglin c with serine proteinases belonging to the subtilisin superfamily (Bode et al., 1987; McPhalen and James, 1988).

The structural organization of the (pro)enzyme-inhibitor contact region is such that the potentially scissile peptide bond cannot move closer to the enzyme nucleophilic group (Ser195 OG atom in proteinases belonging to the chymotrypsin superfamily, and Ser221 OG atom in those belonging to the subtilisin superfamily) due to intramolecular constraints present in the inhibitor molecule (Huber and Bode, 1978; Read and James, 1986; Bode and Huber, 1991, 1992; Stubbs and Bode, 1994). Their modification (i.e. the depletion of the salt bridges supporting the eglin c reactive site), through site-directed mutagenesis, allows tighter enzyme-inhibitor interactions at the catalytic centre, and turns eglin c into a poor serine proteinase inhibitor or substrate (Heinz et al., 1991, 1992).

The tight binding and structural complementarity of the contact interface between the (pro)enzyme and the inhibitor do not allow the diffusion of solvent water molecules to the proteinase catalytic centre, inhibiting the release, through a deacylation process, of the inhibitor molecule which might become nicked at the P,-P,’ reactive site bond (Huber and Bode, 1978; Read and James, 1986; Bode and Huber, 1991, 1992; Heinz et al., 1991, 1992; Stubbs and Bode, 1994).


Site-directed mutagenesis experiments have been performed on eglin c in order to test the mechanism of serine proteinase inhibition and to provide inhibitors endowed with modified specificities. In this respect, amino-acid substitutions introduced on eglin c by Heinz et al. (1991, 1992), aimed at two main goals: (1) modification of the surface complementarity between the inhibitor and the target proteinase; and (2) perturbation of the reactive site loop internal stability and conformation.

In the single Leu45 + Arg eglin c mutant, the presence of the Arg residue at the P, position provides an excellent bovine l3-trypsin inhibitor based on the eglin c scaffold (see Table lo), in agreement with the conserved inhibitor reactive site loop ideal conformation, and with the charge compensation between the eglin c Arg45(P,) residue and the Asp189 side-chain present at the S, specificity subsite of bovine l3-trypsin. Moreover, the Leu45 --) Arg eglin c mutant is still a good B. subtilis subtilisin inhibitor (see Table lo), reflecting the broad enzyme specificity, but does not inhibit thrombin activity, due to the steric clash with the rims of the serine proteinase narrow active site cleft (Heinz et al., 1991, 1992; Stubbs and Bode, 1994).

The single Thr44 -+ Pro eglin c mutant shows a lower affinity for human leukocyte elastase as compared to the wild-type inhibitor (see Table 10). This behaviour has been related to loss of the eglin c intramolecular hydrogen bond between Thr44(P,) and Arg53 residues, which leads to a lower stabilization of the binding loop conformation. This observation is supported further by the loss of human leukocyte elastase inhibition by the single Arg53 += Lys eglin c mutant, which also weakens the Thr44-Arg53 intramolecular interaction, possibly leading to moderate conformational changes in the inhibitor reactive site region. In this respect, it should be noticed that whereas the single Arg53 + Lys eglin c mutant is a B. subtilis subtilisin inhibitor, the single Thr44 -_, Pro eglin c mutant is a substrate (see Table 10) (Heinz et al., 1991, 1992).

The double Leu45 + Arg-Asp46 -+ Ser eglin c mutant is expected to alter both the serine proteinase specificity as well as the intramolecular stabilization of the inhibitor reactive site conformation and rigidity. In fact, the absence of the Asp residue at the P,’ site results not only in the loss of electrostatic interaction between this site and the Arg51 residue, but also in the disruption of a peculiar intramolecular hydrogen bond observed between the backbone amide group of residue 46 arid its own side-chain (Bode et al., 1986, 1987; Heinz et al., 1992). Indeed, multidimensional nuclear magnetic resonance experiments have demonstrated that free wild-type eglin c and the Leu45 + Arg-Asp46 + Ser double mutant have virtually identical conformations (Heinz et al., 1992). However, in the eglin c double mutant, which is a substrate for bovine P-trypsin, human leukocyte elastase and B. subtilis subtilisin, the scissile peptide bond demonstrates much higher mobility than in the wild-type inhibitor (Heinz et al., 1992).

As a whole, the site-directed mutagenesis experiments on eglin c (Heinz et al., 1991, 1992) indicate that subtle modifications in the hydrogen-bonding pattern formed at the proteinase binding loop, both intramolecularly as well as at the enzyme interface, can determine the substrate versus inhibitor fate of eglin c mutants. Moreover, these data indicate that the modulation of the dynamic properties of eglin c may also play a central role in the protein-protein recognition process.


Functional Aspects

Eglin b and c obey the ‘standard mechanism’ of inhibition (see Scheme 2) both from the thermodynamic and kinetic viewpoints (Fink et al., 1986a; Seemtiller et al., 1977, 1986; Braun et al., 1987; Faller et al., 1990; Ascenzi et al., 1991; Junger et al., 1992; Heinz et al., 1992; Cutruzzola’ et al., 1993).

The inhibitory properties of native eglin b and c, differing at position 35, of recombinant N-a-acetylated eglin c, and of synthetic eglin c are undistinguishable (see Table 8) (Seemtiller et al., 1977, 1986; Okada et al., 1990a; Ascenzi et al., 1991). Chemical modification of Tyr residues selectively decreases the inhibitory activity of eglin c, but not of eglin b, whereas modification of His residues selectively decreases the activity of eglin b (Seemtiller et al., 1986). Selective oxidation of both Tyr and His residues decreases the eglin c inhibitory activity, which is inactivated by exhaustive oxidation (Smith et al., 1987; D ean et al., 1989). These observations agree with the fact that the chemically-modified residues Tyr49 and His68, of eglin b and c, are in contact with the target enzyme; moreover, the conformation of the inhibitor reactive site loop may be partly altered due to modification of His and Tyr residues (Seemtiller et al., 1986; Smith et al., 1987).

Eglin c binds to the leucine-specific serine proteinase from spinach (Spinacia olerucea L.) leaves (Leu-proteinase) with the highest known affinity (i.e. K value) for the serine (pro)enzyme-eglin c complex formation (see Table 9). Such a behaviour probably reflects the selective specificity of Leu-proteinase for substrates and inhibitors (such as eglin c) containing a leucyl residue at their P, position (Ascenzi et al., 1991).

According to the serine proteinase primary specificity and to the inhibitor recognition mechanism, eglin c displays a similar affinity for the homologous serine proteinases human cathepsin G, rat cathepsin G, human leukocyte elastase, rat leukocyte elastase, bovine a-chymotrypsin and human pancreatic elastase (see Table 9). all acting pre- ferentially on substrates and inhibitors containing aromatic or bulky aliphatic amino-acid residues at their P, position (Ascenzi et al., 1991). Alternatively, the affinity of eglin c for the cY,-macroglobulin-bound human leukocyte elastase is lower than that observed for the free serine proteinase by approximately 5000-fold (Stack et al., 1988). Moreover, the ability of eglin c to control the catalytic activity of human leukocyte elastase complexed with elastin is impaired (Morrison et al., 1990).

The affinity of eglin c for bovine cx-chymotrypsin bearing a chemically-modified methionine sulphoxide residue at position 192 (bovine Met(O)192 cr-chymotrypsin) is lower than that for the native enzyme (see Table 9), reflecting an impaired packing at the oxidized enzyme-inhibitor molecular interface (Cutruzzola’ et al., 1993).

The affinity of eglin c for porcine chymase, a chymotryspin-like serine proteinase present in mast cells, is lower than that observed for the formation of the bovine a-chymotrypsin-inhibitor complex (see Table 9). This behaviour possibly reflects structural differences occuring at the enzyme secondary recognition subsites (Fink et al., 1986a; Fukusen et al., 1987).


In spite of the lack of structural homology between the different serine proteinase superfamilies, eglin c binds to human cathepsin G, rat cathepsin G, human leukocyte elastase, rat leukocyte elastase, human pancreatic elastase and bovine ol-chymotrypsin with the same affinity found for B. subtih subtilisin, B. licheniformis subtilisin and thermitase (see Table 9). Indeed, although a quantitative evaluation is complex, and should also take into account entropic effects related to solvent displacement/ immobilization occurring upon complex formation, the intermolecular interactions observed in the subtilisin- and thermitase-eglin c complexes are comparable with those observed in the inhibitor adducts with mesentericopeptidase and bovine ol-chymotrypsin (Ascenzi et al., 1991; Bode and Huber, 1991, 1992; Frigerio et al., 1992; Heinz et al., 1992; Krystek et al., 1993).

The affinity of eglin c for bovine ol-chymotrypsin is higher than that reported for porcine pancreatic elastase- and for bovine p-trypsin-inhibitor complex formation (see Table 9), in agreement with the different primary specificities of these homologous serine proteinases (Ascenzi et al., 1991; Heinz et al., 1992). Accordingly, eglin c does not inhibit the trypsin-like serine proteinases bovine thrombin, porcine pancreatic kallikrein and porcine plasmin, acting on cationic substrates (Seemtiller et al., 1977).

The affinity of eglin c for bovine cw-chymotrypsinogen A is lower than that reported for the active enzyme (see Table 9). Such a behaviour can be related to the poor structured state of the activation domain of the zymogen as compared to that of active bovine a-chymotrypsin. In fact upon eglin c binding, a fraction of the free energy for complex formation is used to shape the activation domain of bovine cr-chymotrypsinogen A into a more rigid and defined conformation (Ascenzi et al., 1991).

Eglin c can also form a binary complex with human leukocyte proteinase3, horse leukocyte elastase, porcine leukocyte elastase, bovine anhydrochymotrypsin, mesentericopeptidase, subtilisin DY, and a membrane-bound metallo-endopeptidase from rat kidney hydrolyzing parathyroid hormone. Alternatively, eglin c does not inhibit serine (pro)enzymes of the vital clotting, fibrinolysis and complement cascade, and elastin degradation catalyzed by human and rat metalloelastase secreted by alveolar macrophages (Seemtiller et al., 1977, 1986; Senior et al., 1989; Rao et al., 1991; Dauter et al., 1991; Wiedow et al., 1991, 1993; Yamaguchi et al., 1991; Betzel et al., 1993).

Values of the affinity constant K for eglin c binding to serine (pro)enzymes obtained experimentally (see Table 9) are in excellent agreement with those obtained from empirical free energy calculations based on X-ray crystallographic structures (Krystek et al., 1993). Moreover, as reported for several (pro)enzyme/inhibitor systems, eglin c binding to serine proteinases belonging to the chymotrypsin and subtilisin superfamilies and bovine a-chymotrypsinogen A is an entropy-driven process (Ascenzi et al., 1991).

None of the eglin mutants reaches the strong inhibitory potential of wild-type eglin c against human leukocyte elastase. Eglin c mutants Thr44 --+ Pro, Arg51 + Lys and Arg53 + Lys are lo- to lOO-fold weaker inhibitors, and mutant Leu45 + Arg demonstrates even weaker inhibition (see Table 10). Alternatively, the double mutant Leu45 + Arg-Asp46 + Ser is a substrate of human leukocyte elastase (Heinz et al., 1992).


According to the enzyme primary specificity (Bode and Huber, 1991, 1992), the affinity of the Leu45 -_) Arg eglin c mutant for bovine S-trypsin is comparable to that of the native inhibitor for chymotrypsin- and subtilisin-like serine proteinases (see Tables 9 and 10). Alternatively, the eglin c double mutant Leu45 + Arg-Asp46 --, Ser is a substrate of bovine P-trypsin (see Table 9) (Heinz et al., 1992). Besides wild-type eglin c, only the mutants Leu45 + Arg and Arg53 + Lys inhibit B. subtilis subtilisin. On the other hand, mutants Thr44 -+ Pro, Leu45 + Arg-Asp46 + Ser and Arg51 + Lys are substrates of B. sub&is subtilisin (see Table 10) (Heinz et al., 1992).

Data reported in Table 10 indicate that point mutations may affect serine proteinase- inhibitor complex stability, and may transform a protein inhibitor into a substrate, possibly reflecting the increased flexibility of the inhibitor reactive site loop (Heinz et al. , 1992).

Eglin c present in the complex with B. licheniformis subtilisin and bovine a- chymotrypsin was found to be shortened at the N-terminus by seven amino-acid residues (Thrl-Glu2-Phe3-Gly4-SerS-Glu6-Leu7) (Bode et al., 1986, 1987; Dauter et al., 1991; Frigerio et al., 1992). Modification of eglin c at the N-terminus by proteases from Nocurdiu mediterrunei has also been reported (Schar et al., 1988). Moreover, multiple hydrolytic cleavages of the N-terminal chain have been observed during purification and crystallization of the eglin c homologous chymotrypsin inhibitor CI-2 from barley seeds (McPhalen and James, 1987). The Ser5-Gly70, Leu7-Gly70 and Lys8-Gly70 eglin c derivatives display the same affinity as the native inhibitor for human cathepsin G, human leukocyte elastase and bovine ol-chymotrypsin (see Table 11) (Dodt et al., 1987; Okada et al., 1990b). However, deletion of the SerS-Glu6 dipeptide from the relatively heat stable Ser5-Gly70 eglin c leads to a heat labile Leu7-Gly70 inhibitor. This suggests that the Thrl-Lys8 N-terminal domain has no direct influence on the proteinase-eglin c complex formation, but may be involved in stabilizing the inhibitor native conformation against denaturation (Dodt et al., 1987).

The affinity of peptide fragments of eglin c for human cathepsin G, human leukocyte elastase and bovine cx-chymotrypsin is lower than that observed for the native inhibitor (see Table 11). The low affinity of peptide fragments of eglin c containing the inhibitor reactive loop (i.e. the Ser41-Tyr49 region) for human cathepsin G, human leukocyte elastase and bovine cY-chymotrypsin has been related to the fact that these peptides do not reach sufficient intramolecular stabilization to maintain the correct and rigid conformation required for the productive inhibitory interaction with the target enzyme (Okada et al., 1989, 1990b).

Values of the second-order rate constant for the serine proteinase-eglin c complex formation (i.e. k,, values) range between 1.1X103 and 2.5X 107 M-l set-l (see Table 9). The value of k,, for eglin c binding to the cr,-macroglobulin-bound human leukocyte elastase is approximately lO,OOO-fold lower than that for the free serine proteinase (Stack et al., 1988). Moreover, heparin reduces the k,, value for eglin c binding to human leukocyte elastase by approximately 200-fold (Frommherz et al., 1991). Values of the first-order rate constant for the enzyme-inhibitor complex dissociation range between 1.2x10-4 and 1.1x10-i set-1 (see Table 9). This behaviour indicates that the accessibility of eglin c to the serine proteinase active centre (expressed by


k,, values) depends on the enzyme considered. Moreover, kinetics of the serine proteinase-eglin c complex dissociation (expressed by values of k,,,) indicate that polypeptide loop(s) protruding into the (pro)enzyme active site area and strongly contacting the inhibitor should contribute to the binary complex destabilization (Ascenzi et al., 1992; Cutruzzola’ et al., 1993).

As reported for several (pro)enzyme/inhibitor systems, the formation of the stable inactive human leukocyte elastase-, bovine a-chymotrypsin-, bovine Met(O)192 (Y- chymotrypsin- and porcine chymase-eglin c binary complexes involves two steps: a fast binding reaction followed by a slow rate-limiting process (Fink et al., 1986a; Braun et al., 1987; Cutruzzola’ et al., 1993). Alternatively, the human pancreatic elastase-eglin c complex (de)stabilization follows a simple behaviour (Faller et al., 1990).

Eglin c binding to serine proteinases and bovine or-chymotrypsinogen A reflects the acidic pK shift of the invariant His catalytic residue (at position 57 and 64 in the serine (pro)enzymes belonging to the chymotrypsin and subtilisin superfamilies, respectively) from 6.9, in the free (pro)enzyme, to 4.5, in the serine (pro)enzyme%glin c complex. In this respect, the observed acidic pK shift of the one-proton binding group could be interpreted as reflecting a change in the dielectric constant at the (re)active site, and the strengthening of the serine (pro)enzyme hydrogen bond between Ser195 OG and His57 NE2 atoms in serine (pro)enzymes belonging to the chymotrypsin superfamily, and between Ser221 OG and His64 NE2 atoms in B. licheniformis subtilisin, upon complex formation. In analogy with other serine (pro)enzyme/macromolecular inhibitor systems, such hydrogen bonding, occurring upon binary adduct formation, is very weak or absent in the free proteinases and their zymogens (Huber and Bode, 1978; Bolognesi et al., 1982; Ascenzi et al., 1991).

Eglin c binding to human neutrophils is a saturable, time-dependent and largely reversible process, with a K value of 5.0~10s M-l, and a t,,, value of about 20 min (pH 7.3 and 5°C). Eglin c dissociates from cells, though in a fashion not strictly in accord with a simple first-order process. This behaviour may indicate heterogeneity among the binding sites or partial internalization. Approximately 100,000 eglin c binding sites per cell have been estimated. Eglin c binding to human neutrophils is essentially unaffected by other serine proteinase inhibitors, indicating that the eglin c binding site may not be an active serine proteinase, such as cathepsin G, elastase or proteinase 3. Alternatively, the affinity of eglin c for cells is enhanced 4-fold by the chemotactic peptide N-formyl-Met-Leu-Phe. This behaviour indicates that the interaction of the two compounds with human neutrophils is not mutually exclusive. Much higher levels of eglin c (1.7x10-7 and 2.5x10-5 M) are required in order to inhibit the generation of oxygen free radicals in human neutrophils, and the N-formyl-Met-Leu-Phe induced chemotaxis of rat neutrophils, respectively (Braun and Schnebli, 1987; Rigbi et al., 1987b). In this respect, eglin c shows a superoxide dismutase-like effect, inhibiting the reduction of cytochrome c in the xanthine-xanthine oxidase system (Suter and Chevallier, 1988). Nevertheless, eglin c does not affect bacterial killing by human polymorphonuclear leukocytes (Esposito et al., 1988).


Biomedical Aspects

Eglin c is a potential therapeutic agent for the treatment of disease states associated with inflammation. In particular, eglin c has been proven effective for the treatment of shock and emphysema in experimental animal models. Eglin c has been found to be well tolerated, with no significant effects on the cardiovascular and central nervous systems, basic metabolism, clotting, fibrinolysis and complement systems. The main safety concern, however, relates to the allergenic potential of the inhibitor (Braun and Schnebli, 1986; Jochum et al., 1986; Schnebli and Braun, 1986; Seemtiller et al., 1986; Siebeck et al., 1992a, b). Being a protein, eglin c can only be administered to animals or patients intravenously, intratracheally, intraperitoneally or topically. Inhibitor concentrations in lymph were comparable to those in plasma. Eglin c is eliminated rapidly from circulation through the kidneys into urine after intravenous administration in rodents and the pig (with a half-life of approximately 10 min and 2 hr, respectively); therefore, plasma levels can be closely controlled (Braun and Schnebli, 1986; Schnebli and Braun, 1986; Seemtiller et al., 1986; Nick et al., 1988; Teschner et al., 1988; Goddard et al., 1991; Siebeck et al., 1992a, b).

Considering the central role played by elastase in the process of leukocyte infiltration and accumulation in inflamed microvessels, eglin c could be used to prevent neutrophil infiltration (adhesion, penetration and migration) into inflamed vessels and neutrophil- mediated injury to the microvascular endothelium. The fact that eglin c significantly attenuates these processes in vivo suggests that this serine proteinase inhibitor may be useful in controlling pulmonary septic shock (Inauen et al., 1990; Zimmerman and Granger, 1990; Siebeck et al., 1992a, b; Weis, 1992; Pintucci et al., 1993; Weng, 1993; Woodman et al., 1993; Yamamoto et al., 1993; Hansen and Stawski, 1994; Totani et al. , 1994).

The therapeutical application of ‘eglin c in the treatment of septic lung oedema induced in pigs, by gram negative endotoxin and Escherichia coli bacteremia, has indicated that alterations in the systemic and pulmonary circulation were in part prevented by administration of the inhibitor. Thus, eglin c significantly reduces the loss of intravascular protein and plasma concentration of leukocyte elastase, without affecting systemic hypotension (Siebeck et al., 1992a, b; Weis, 1992; Weng, 1993). In this respect, eglin c inhibits human leukocyte elastase which increases permeability of cultured pulmonary endothelial cell monolayers (Suttorp et al., 1993), and the elastolytic activity present in the bronchial secretions from patients with cystic fibrosis (Suter et al., 1986). Nevertheless, elastolytic proteinases from Staphylococcw aurezu and Pseudomonas aeruginosa may inactivate eglin c (Sponer et al., 1991).

In the rat, eglin c prevents the oedematous injury in isolated lungs (Baird et al., 1986), and attenuates the mucosal permeability in the terminal ileum (von Ritter et al., 1989). Nevertheless, peritoneal rat macrophages may inactivate eglin c (Dean and Schnebli, 1989). Moreover, this serine proteinase inhibitor affects pulmonary antimicrobial mechanisms important for preventing and eradicating bacterial infection of the lower respiratory tract, in the mouse (Esposito et al., 1987).


Alternatively, eglin c is ineffective in healing endotoxin shock in sheep, possibly reflecting the low second-order combination rate constant for inhibitor binding to ovine leukocyte elastase (see Table 9) (Junger et al., 1989; 1992).

Therapeutic effects of eglin c in endotoxin shock of the pig are enhanced if hirudin is also administered. In fact, hirudin reduces the endotoxin-induced fibrinogen consumption and thus the formation of fibrin monomer, as well as the pulmonary vasoconstriction in pigs (Siebeck et al., 1989; Weis, 1992; Weng, 1993).

Considering the manifold involvements of chymase in the allergic inflammatory reaction, eglin c may be a promising candidate for therapeutic approaches also in this field (Fink et al., 1986a; Fukusen et al., 1987).

The discovery of homozygous a,-proteinase inhibitor deficiency and its relation to emphysema, and the key role of elastolysis in experimental serine proteinase induced emphysema, has suggested that elastase-antielastase imbalance in the lung was the root cause of emphysema. It seems therefore rational, considering the need for new emphysema prevention initiatives, to explore a possible role for antielastase agents (Karlinsky and Snider, 1978; Heidtmann and Travis, 1986).

When eglin c is given intratracheally to hamsters 1 hr prior to human leukocyte elastase and human leukocyte proteinase-3 administration, it can prevent or ameliorate the serine proteinase-induced emphysema only if present in large molar excess (20-fold) over the 1:l stoichiometry required for the in vitro inactivation of the serine proteinases (Snider et al., 1985; Kao ef al., 1988). This observation, together with the inhibitor short half-life in the lung (about 2 hr) (Lucey et al., 1986) makes this eglin c administration route for the treatment of emphysema impractical. Thus, since eglin c has an extremely short circulating half-life in rodents (about 10 min) after intravenuous administration, a conjugate of the inhibitor with poly(oxyethylene) (>20 kDa average molecular weight) has been prepared. The modification of eglin c with poly(oxyethylene) does not impair the inhibitor ability to bind leukocyte elastase; moreover, higher plasma concentration can be achieved in comparison with the free inhibitor (Goddard et al., 1991).

Histological examination of the lungs has confirmed that eglin c is highly protective against the effects of instilled leukocyte elastase (Schnebli and Braun, 1986). As expected (see Table 9), the porcine pancreatic elastase induced emphysema in rats is unaffected by eglin c (Lai and Diamond, 1990). Nevertheless, eglin c does not suppress the vascular leakage induced in hamster cheek by application of leukotriene B4 (Rosengren and Arfors, 1990), and the basement membrane invasion by bovine capillary endothelial cells induced by the basic fibroblast growth factor during in vitro angiogenesis (Mignatti et al., 1989).

Eglin c also inhibits: (1) degradation of proinsulin/insulin by serum pancreatic serine proteinases from acute pancreatitis patients (Lau et al., 1990); (2) platelet activation, platelet aggregation, and serotonin and plasminogen activator inhibitor-l release by cathepsin G (Renesto and Chignard, 1991; Renesto ef al., 1990, 1991; Maugeri et al., 1992; Pintucci et al., 1993); (3) h uman leukocyte proteinase-3 activity on platelets (Renesto et al., 1994); (4) platelet reaction to shear-stress and collagen, and


spontaneous thrombolysis (Yamamoto et al., 1993); (5) leukocyte elastase-mediated release of the von Willebrand factor from cultured endothelial cells (Chignard et al., 1993); (6) inactivation of cr,-macroglobulin by activated neutrophils in arthritic joints (Abbink et al., 1991; Steinmeyer and Kalbhen, 1991); (7) inactivation of endothelin by polymorphonuclear leukocyte-derived lytic enzymes (Patrignani et al., 1991); (8) neutrophil-mediated renal injury, in rats (Linas et al., 1991, 1992); (9) porcine leukocyte elastase and cathepsin G-induced release of aortic endothelium-derived growth factors (Totani et al., 1994); and (10) degradation of elastin-fibrin biomaterial induced by human pancreatic elastase (Collet et al., 1991).

Eglin c may interfere with the activation of tumour cell urokinase. In fact, human leukocyte elastase inactivates prourokinase and thereby prevents tumour cell proliferation. Therefore, serine proteinase inhibition may be a drawback in eglin c therapy (Schmitt et al., 1989; 1991; Kanayama and Terao, 1990).

Although the increased release of proteinases from polymorphonuclear leukocytes has been indicated to take part in the pathogenesis of enhanced muscle protein breakdown in renal failure, eglin c does not decrease the catabolism in uremic rats (Teschner et al., 1988). Moreover, eglin c does not affect: (1) the plasminogen activator involved in hCG-induced neutrophil extravasion and vasopermeability increase in the rat testis (Loukusa et al., 1990); (2) complement activation during storage of preserved blood (Schleunig et al., 1990; 1992); and (3) eosinophil peroxidase release (Kawaji et al., 1993).

Finally, eglin c does not induce alterations in lung, liver and kidney functions (Schnebli and Braun, 1986).

Chapter 6

Bdellin

General Aspects

Inhibitors of trypsin, plasmin and sperm acrosin were first discovered by Fritz et al. (1969) in crude preparations of hirudin which had been obtained by extraction from whole medicinal leech Hirudo medicinalis. They were named bdellins after the Greek word for the leech (Fritz et al., 1969; Fink et al., 1986b; Seemiiller et al., 1986).

Bdellins occur in all parts of the leech, but the highest concentration is found in the region of the outer sex organs. Therefore, it has been proposed that bdellins may be involved in reproduction (Seemuller et al., 1986).

Structural Aspects

Two groups, bdellins A and B, each consisting of several closely related members, have been distinguished on the basis of their behaviour in anion-exchange chromatography, and numbered in order of elution, A-l to A-5, and B-l to B-6 (Seemtiller et al., 1986).

Bdellins A and B differ significantly in their amino-acid composition, but only minor differences exist between the amino-acid compositions of the individual member of each group. The molecular mass is 7 kDa for the A-type and 5 kDa for the B-type bdellins. Moreover, a high molecular mass form of bdellin B-3 (20 kDa; HMB B-3) has been characterized. The polypeptide chain of the HMB B-3 inhibitor contains the bdellin B-3 domain and is elongated at the C-terminus by an extension peptide which constitutes about two-thirds of the total molecule. Under non-denaturating conditions, HMB B-3 may associate forming dimers or even trimers (Fink et al., 1986b; Seemuller et al., 1986).

Bdellin B-3, represented by the N-terminal 46 residues of HMB B-3, is a non-classical Kazal-type inhibitor with respect to the position of the 6 half cystines, demonstrating 44% amino-acid identity with leech-derived tryptase inhibitor (LDTI-A) (see Fig. 11) (Fink et al., 1986b; Seemtiller et al., 1986; Menegatti et al., 1987b; Sommerhoff et al. ) 1994).

275

276

LDTI-A Bdellin B-3 Rhodniin domain 1 Rhodniin domain 2 PST1 OMSVP3




LDTI-A Bdellin B-3 Rhodniin domain 1 Rhodniin domain 2 PST1 0MSVP3


P. Ascenzi et al.

Lys-Pro.Val His.Arg.Val His.Arg.Val Lys.Pro.Val Asn.Pro.Val Arg.Pro.Leu

AlanSex

Gly(AsnfLys

r

Fig. 11. Amino-acid sequences of the leech-derived tryptase inhibitor (LDTI-A), of bdellin B-3, of rhodniin domains 1 and 2, of the porcine pancreatic secretory trypsin inhibitor (PSTI) and of the silver pheasant ovomucoid third domain (OMSVP3) (Fink et al., 1986b; Menegatti et

al., 1987b; Sommerhoff et al., 1994). The arrow marks the P,-P,’ reactive site bond. Gaps are indicated by dashes. Identical residues are boxed. Multiple alignments were produced using the Genetics Computer Group sequence analysis software package (GCG version 7.1) using a

Vax/VMS system (Devereux et al., 1984).

A molecular model of bdellin B-3 has been built, based on sequence homology and on the known tertiary structure of Kazal-type inhibitors (Fink et al., 1986b). In bdellin B-3 tertiary structure organization, the central scaffold consisting of a a-helical segment (Asn33_Gly41), a three-stranded antiparallel P-sheet, the VallSHis21 binding loop, and the Cys16-Cys35 and Cys24-Cys56 disulfide bridges appear to have been basically conserved. The reactive bdellin B-3 binding loop, linked to the core through the Cysl6-Cys35 disulfide bridge and through residue Asn33, is expected to maintain a ‘canonical’ structure, allowing tight association to trypsin (Fink et al., 1986b). In bdellin B-3 (HMB B-3 included), the following residues comprise the reactive site loop: VallS(P,), Cysl6(P,), Thrl7(P,), Lysl8(P,), Glul9(P,‘), Leu20(P,‘) and His21(P,‘).


Lysl8, at position P,, is the primary determinant of the inhibitor specificity, consistent with the trypsin-like serine proteinase inhibition pattern (Fink et al., 1986b).

It is not clear as yet, whether the different forms are derived from single molecular species of bdellins A and B, whether they represent genetic variants in the leech population, or whether each individual leech has a family of genes for each bdellin. Moreover, it is not known whether low molecular mass bdellin B-3 is released from HMB B-3 (or from a precursor of HMB B-3 with even higher molecular mass) in the leech under physiological conditions, or whether it is cleaved off in the course of isolation (Fink et al., 1986b; Seemtiller et al., 1986).

Functional Aspects

Bdellins A and B as well as HMB-B-3 obey the ‘standard mechanism’of inhibition (see Scheme 2) both from the thermodynamic and kinetic viewpoints (Fink el al., 1986b; Seemtiller et al., 1986). Bdellins A apd B are strong inhibitors of trypsin, plasmin and sperm acrosin, with values of the inhibition constant (K) ranging between 107 and 1010 M-‘. Alternatively, bdellins A and B do not affect the catalytic properties of bovine ol-chymotrypsin, tissue and plasma kallikreins, as well as subtilisins (Seemtiller et al., 1986).

The inhibitory properties of bdellin B-3 in vitro are not influenced by the C-terminal extension peptide. In this respect, the value of the affinity constant for bdellin B binding to boar sperm acrosin (K =lOlO M-l) is consistent with the value of K (1.3X 109 ~-1)

estimated for HMB B-3 association. It is possible that the C-terminal extension peptide plays a role at physiological level, for example, by attaching the inhibitor to special molecules or membranes (Fink et al., 1986b; Seemtiller et al., 1986).

Biomedical Aspects

A possible therapeutic use of bdellins, especially of bdellin A, could be as a plasmin inhibitor to control bleeding. In this respect, it is important to note that bdellins, administered in a systematic way, are rapidly excreted into the urine (Seemtiller et al. , 1986).

Chapter 7

Ttyptase Inhibitor

General Aspects

Human mast cells contain in their secretory granules high amounts of tryptase, a trypsin-like serine proteinase, in a tetrameric, proteoglycan-associated form. Although the physiological function of human tryptase still remains to be elucidated, pathogenetic roles of this serine proteinase have been discussed in several diseases, such as allergy, asthma, rheumatoid arthritis, fibrosis, artherosclerosis, scleroderma, anaphylaxis and mastocytosis. Moreover, the membrane-associated tryptase TL,, located on human T4+ lymphocytes, seems to be involved in an early phase of docking of the HIV-l virus on lymphocytes and on HIV-l infection as well. In vitro studies suggest that human tryptase is also involved in the catabolism of extracellular matrix proteins, in coagulation, and in neuropeptide turnover (Eklund and Stevens, 1993). Human tryptase is virtually unique among the serine proteinases as it is fully catalytically active in plasma and in the extracellular space. The enzyme action is unaffected by the human naturally occurring antiproteinases (such as a,-proteinase inhibitor, antithrombin III, C,-esterase inhibitor, mucus proteinase inhibitor and a,-macroglobulin), regulating the activity of other trypsin-like serine enzymes. Among serine proteinase inhibitors derived from non-human species or produced by recombinant technologies, the so-called leech-derived tryptase inhibitor (LDTI) is able to modulate human tryptase activity (Auerswald et al., 1994; Sommerhoff et al., 1994). Moreover, the activity of the membrane-associated tryptase TL, located on human T4+ lymphocytes is specifically abolished by trypstatin, a physiological inhibitor of rat tryptase isolated from rat mast cells (Eklund and Stevens, 1993). Considering that mast cell proteinases are released during the parasitic infections, it is conceivable that LDTI is secreted by the leech Hirudo medicinalis to block host defence mechanisms initiated by the mast cell enzymes (Auerswald et al., 1994; Sommerhoff et al., 1994).

Furthermore, it is known that the bovine basic pancreatic trypsin inhibitor (BFTI) affects the catalytic properties of tryptase from bovine mast cells (Fiorucci et al., 1995).

279

280

Structural Aspects

P. Ascenzi et al.

Three forms of LDTI, differing at the C-terminus, have been identified. LDTI-A (42 amino acids), the smallest non-classical (with respect to the position of the characteristic six half cystines) Kazal-type inhibitor described so far, is shorter than LDTI-B, which contains Gly43 as additional C-terminal residue. LDTI-C (46 amino acids) displays the additional C-terminal Ile44-Leu45-Asn46 sequence (Sommerhoff et al., 1994). LDTI contains six invariant cysteine residues, involved in three disulfide bridges: Cyst-Cys29, Cys6-Cys25, and CyslO-Cys40 (Mtihlhahn et al., 1994).

Sequence comparison revealed a high degree of similarity of LDTI to Kazal-type inhibitors of blood-sucking ectoparasites, such as bdellin B-3 (44% amino acid identity) (Fink et al., 1986b), and domain 1 and 2 of rhodniin (36 and 35% amino-acid identity, respectively), a thrombin-inhibitor from the insect Rho&&s prolixus, the vector of Chagas’ disease (Friedrich et al., 1993) (see Fig. 11).

The topology of the global fold of the recombinant LDTI-C and its secondary structure elements, determined in solution by NMR techniques (Mtihlhahn et al., 1994), are shown in Fig. 12. The core of LDTI-C consists of a short 3,,-helix (one and a half-turn) followed by extended polypeptide structure between Asn30 and Cys40. A fragment from Va113 to Ala22 may be described as a two-stranded antiparallel mini-P-sheet with a reverse turn region at residues GlylS-Argl9 (Miihlhahn et al., 1994).

The N-terminal region of recombinant LDTI-C, which is connected to the core helix via the Cys&Cys29 and Cys6-Cys25 disulfide bridges and contacts at residue Ile9, contains the Cys6Phel2 inhibitor reactive site loop (Mtihlhahn et al., 1994). The geometry of this region of recombinant LDTI-C corresponds to the ‘canonical’ inhibitory loop conformation, comprising the P,-P,’ residues surrounding the scissile LysS-Ile9 bond, enabling binding to the target enzyme (Bode and Huber, 1991, 1992; Mtihlhahn et al., 1994). In LDTI-C, the following residues comprise the canonical reactive site loop: Cys6(P,), Pro7(P,), Ly&(P,), Ile9(P,‘), LeulO(P,‘) and Lysll(P,‘). Lys8, at position Pi, is the primary determinant of the inhibitor specificity, consistent with trypsin-like serine proteinase inhibition (see Table 12). Although conformationally disordered, the Lys8(P,) residue is exposed and poised for insertion into the enzyme primary specificity subsite (i.e. the S, cleft) (Mtihlhahn et al., 1994).

Fig. 12. Ribbon drawing of LDTI-C showing the secondary structure elements and disulfide bridges. The Ly&(P,)-Ile9(P,‘) reactive site bond is in the upper part of the picture. Reproduced

with permission, from Miihlhahn et al. (1994).


The structure of the human cw-thrombin-rhodniin complex has been solved recently (A. van de Locht and W. Bode, personal communication; Stubbs and Bode, 1994), and indicates that contacts are made by both the inhibitory domains of rhodniin along the enzyme active cleft.

Functional Aspects

LDTI obeys the ‘standard mechanism’ of inhibition (see Scheme 2) both from the thermodynamic and kinetic viewpoints (Auerswald et al., 1994; Sommerhoff et al.,

1994). LDTI is the first inhibitor of human mast cell tryptase identified in leeches, and is also a tight-binding inhibitor of trypsin, chymotrypsin, plasmin, tissue kallikrein, thrombin and neutrophyl cathepsin G (see Table 12). Wild-type and recombinant LDTI inhibit all these serine proteinases with identical affinity and mechanism (Auerswald et al., 1994; Sommerhoff et al., 1994).

Remarkably, the degree of tryptase inhibition by LDTI depends on the size of the substrate used. This observation suggests that, most probably due to steric hindrance, LDTI binds only to two of the four catalytic subunits of the tryptase tetramer, leaving the active sites of the other two subunits only accessible to small substrates (see Table 12) (Auerswald et al., 1994; Sommerhoff et al., 1994). A similar behaviour has been reported for BPTI binding to bovine tryptase, which associates four inhibitor molecules with affinity constants ranging between 1.2~108 and 2.3~104 M-l, at pH 8.0 and 3O.O”C (Fiorucci et al., 1995). Alternatively, this tetrameric serine proteinase binds four benzamidine molecules with the same affinity (K = 5.5X 104 M-‘, at pH 8.0 and 3O.o”C; Fiorucci et al., 1995). It appears possible that the tryptase isoforms identified by cDNA cloning differ in the reactivity with substrates and inhibitors (Auerswald et al., 1994; Sommerhoff ef al., 1994).

Table 12. Values of thermodynamic and kinetic parameters for LDTI binding to serine proteinases (pH 7.6 and 25.0”(Z)*

Proteinase

Human tryptase Bovine trypsin Bovine chymotrypsin Human plasmin Porcine tissue kallikrein Human thrombin Human cathepsin G Human plasma kallikrein Human Factor Xa Porcine pancreatic elastase Human neutrophil elastase Human urokinase

K (M-l)

7.1 x loq 1.1 x 109 5.0 x 107

2.1 x 106 8.3 x 105 6.8 x 105 6.2 x 105 < 5 x 105 <5x1@ <5x105 <5x105 <5x10=5

*From Sommerhoff et al. (1994); 7The value of K refers to the interaction of LDTI with the two high affinity active sites present in the tryptase tetramer (see text).


Despite the high similarity in sequence to LDTI (see Fig. ll), neither bdellin B-3 nor rhodniin inhibit tryptase. The different inhibitor binding behaviour may be related to differences in the flexibility observed at both C- and N-termini. In fact, the termini are located near the inhibitor reactive loop, and so may interact in different ways with tryptase or other target enzymes, contributing to the specificity of LDTI. Further specificity may arise from interactions with residues inserted in loops adjacent to the reactive site (Mtihlhahn et al., 1994).

Biomedical aspects

LDTI effectively blocks the tryptase-induced cleavage of the vasoactive intestinal peptide, of the histidine-methionine peptide and of kininogen, representatives of the polypeptides and proteins thought to be biologically-relevant substrates of this proteinase, and also the mitogenic activity of tryptase, an example of the direct cellular effects of tryptase (Auerswald et al., 1994; Sommerhoff et al., 1994).

Recombinant LDTI inhibits HIV-l replication in HUT-78 cells, as already reported for trypstatin and the human urinary trypsin inhibitor (UTI). The observed suppression of HIV-l replication by LDTI should be specific because the inhibition is dose dependent and the tryptase inhibitor demonstrated no cytotoxic effect in the control experiment. Nevertheless, since LDTI does not contain the Gly-Pro-Cys-Arg sequence motif present in UT1 and trypstatin, it remains unclear whether inhibition of HIV-l replication by LDTI follows the reaction pathway suggested for trypstatin, or whether it is due to a different mechanism. In this respect, it may be recalled that trypstatin binds to tryptase TL,, which interacts closely with the external envelope glycoprotein gp120 and the CD4 receptor on T-lymphocytes during virus binding and penetration (Auerswald et al., 1994). Moreover, LDTI may be a useful pharmacological probe to elucidate the role of tryptase in mast cell-related (patho)physiological situations (Sommerhoff et al., 1994).

The tryptase-specific LDTI extends the range of specificities of known leech-derived inhibitors which include: (1) the Factor Xa-specific antistasin and ghilanten; (2) the thrombin-specific hirudin; (3) the tissue kallikrein-specific hirustasin; (4) the tryptase-specific LDTI; (5) decorsin and ornatin, antagonists of glycoprotein IIb-IIIa and inhibitors of platelet aggregation; (6) bdellin, inhibiting plasmin- and trypsin-like enzymes; and (7) eglin, inhibiting chymotrypsin- and subtilisin-like enzymes. Except for the anticoagulants hirudin, antistasin, ghilanten, hircustosin, decorsin and ornatin, which are essential to maintain the liquid state of the ingested blood inside the leech, and LDTI, which may inhibit the host defence mechanisms initiated by the mast cell enzymes, the biological functions of bdellin and eglin are not yet understood (Fink et al., 1986b; Seemtiller et al., 1986; Bode and Huber, 1991,1992; Sommerhoff et al., 1994; Sollner et al., 1994; Stubbs and Bode, 1994).

Chapter 8

Conclusions

The reactivity of macromolecular inhibitors from the leech toward physiologically important proteinases indicates the possibility of various applications including medicinal usage. Therefore, the relevant question can be formulated as follows: “what is needed and what is being done in the near future to develop new proteinase inhibitors for the therapeutic use?” The spontaneous answer is: “a lot”. However, the answer appears to be somewhat cloudy, in that there is a large agreement on the following proposal: the operating plan should consist in “a bank of inhibitors with some human and some from other sources, some large and some small, some antigenic and some tolerated, some oxidizable and/or degradable and some nonoxidizable and nondegradable” (Travis and Fritz, 1991). In other words, an inhibitor for every medical or surgical situation should be available. Thus, for emergency therapy (e.g. in septic shock or after multiple trauma) only one administration will be useful and therefore immunological tolerance could not be a prerequisite for an ideal inhibitor (Travis and Fritz, 1991). Alternatively, in patients with emphysema (where the inactivation of natural inhibitors by oxidants causes a local decrease of inhibitory capacity) (Johnson and Travis, 1978; Janoff, 1985a; Banda et al., 1987; Fritz, 1988) rapidly acting low molecular weight non-oxidizable inhibitors of human neutrophil elastase would be useful during short-term therapy (Travis and Fritz, 1991). Related to the development of new inhibitors as drugs is the issue of antiproteinase selectivity. Thus, many synthetic low molecular weight inhibitors not only interfere with the coagulation cascade, also but compromise fibrinolysis and thrombolysis (Tapparelli et al., 1993); this concern will hopefully be overcome by further optimization of the structure of inhibitors with regard to cognate proteinases (e.g. those related to potential alteration at various corners of the hemostatic balance network), by interacting in an optimized way with the enzyme secondary recognition subsites or, better, with additionally binding clefts outside the active region (Ascenzi et al., 1992b).

Proteinase inhibitors from the European medicinal leech Hirudo medicinalis are good candidates for developing drugs for cardiovascular diseases and severe inflammation (Schnebli and Braun, 1986; Seemtiller et al., 1986; Sollner et al., 1994; Stubbs and Bode, 1994).

283


In animal models and in clinical experience, hirudin and hirulog have been demonstrated to be effective in both venous and arterial thrombotic situations, offering possible advantages over presently available anticoagulants such as heparin (Wallis, 1989; Kelly et al., 1991). Ongoing or planned human clinical trials will evaluate the safety and efficacy of hirudin (and hirudin analogs) in unstable angina, restenosis following coronary angioplasty, as an adjunct to fibrinolytic therapy, for treatment of disseminated intravascular coagulation and possibly antithrombin III deficiency (Johnson, 1994). Many advantages due to the high specificity of proteinase inhibitors from leech are already clear. Thus, by virtue of their mode of action, all anticoagulant drugs cause bleeding at high doses, but subtle differences in their specificity profile (i.e. in the recognition ability) are expected to make some agents less predisposed to side-effects (such as bleeding) than others. This was confirmed in a human volunteer study with hirulog, alone or in the presence of aspirin, where adequate anticoagulation was achieved without prolongation of the bleeding time (Fox et al., 1993).

Several lines of evidence have implicated neutrophil elastase in the development of adult respiratory distress syndrome (Repine, 1992). In general, the data which have been accumulated in the past indicate that in the absence of sufficient regulating proteinase inhibitors, neutrophil enzymes can cause abnormal connective tissue damage. It is estimated that between 0.3 and 0.5 g of neutrophil elastase and 0.5 g of cathepsin G are normally released from phagocytic cells per day in the average normal individual; for those suffering from an inflammatory episode these values rise dramatically, perhaps as much as lo-fold higher (Travis and Fritz, 1991). As a result, due to saturation of natural inhibitors (mainly cw,-proteinase inhibitor, a,-macroglobulin and mucus proteinase inhibitor (Wewers et al., 1988; Virca and Travis, 1990; Travis and Fritz, 1991) by massive proteinase release, a functional deficiency is created where inhibitor levels are normal, but biological activity is at or near zero (Travis and Salvesen, 1983). Therefore, for medical purposes, an effective and specific inhibitor for human neutrophil elastase could be important to be designed for the following reasons. When lysosomal enzymes are secreted into serum, neutrophil elastase represents the most active and the most unspecific proteinase then present, initiating unspecific activation for other proteolytic activities. This leads to loss of blood hemostasis and, via kinin production, to increased vascular permeability and reduced blood pressure (Jochum et al., 1981; Duswald ef al., 1982). Excessive leakage of neutrophil elastase from leukocyte is correlated with various types of tissue damage (Janoff, 1985b). Attempts have been taken to outdo nature by generating non-oxidizable mutants of human inhibitors (e.g. cw,-proteinase inhibitor with valine or leucine replacing methionine at position Pi) (Travis et al., 1985). Several other naturally occurring serine proteinase inhibitors are also being produced by recombinant DNA technology (Auerswald et al., 1988; Maywald et al., 1988; Oltersdorf et al., 1989; Hecht et al., 1991). These include eglin c (Travis and Fritz, 1991): in this case, however, problems will be potential antigenicity and rapid elimination due to the relatively small size of the molecule.

Other proteinase inhibitors from leech offer interesting potentialities for medical application; in particular, antistasin and related molecules (e.g. hirustasin) display possible antimetastatic properties (Sollner et al., 1994).


Finally, macromolecular inhibitors from leech (and cognate proteinases) can be of general relevance in protein-protein recognition, due to their potential utility for understanding the molecular mechanism of enzyme-inhibitor interaction; such knowledge, in fact, will spur new approaches to therapy, since well-designed new inhibitors can be obtained by chemical synthesis (e.g. Tsuda et al., 1994). It is known that minor changes in structure may influence thermodynamics and kinetics of protein-protein interactions in such a way that observed modifications in activity can only be rationalized by three-dimensional modelling (Menegatti et al., 1987b). In other words, the fine-tuning improvement of strong interactions may be difficult to understand without a knowledge of the structural model (Wells and Lowman, 1992). Attempts to illuminate the structural properties confering inhibitory activity to proteins from leech have be successful, due to elimination of uncertainties in the local folding. Thus, the structure of the thrombin-hirudin adduct (Grtitter ef al., 1990; Rydel et al., 1991) has provided unvaluable insight into the action of hirudin. The same is worth for antistasin (Stubbs and Bode, 1994), decorsin (Krezel et al., 1994), eglin c (e.g. Frigerio et al., 1992), and of the leech tryptase-derived inhibitor (Mtihlhahn et al., 1994). The diverse approaches to achieve proteinase inhibition nature has been used and have been illustrated (Bode and Huber, 1992).

Acknowledgements

Authors thank Dr Giuseppina Mignogna for helpful discussions on amino-acid sequences of leech proteinase inhibitors, and Dr Stefano Santanche for help in reference search and check.

This study has been supported by the Italian Ministry of University, Scientific Research and Technology (MURST), as well as by the Italian National Research Council (CNR, target oriented projects Biotecnologie e Biostrumentazione, Chimica Fine II and Invecchiamento, as well as the special project Peptidi Bioattivi).

References

Abbink, J. J., Kamp, A. M., Nieuwenhuys, E. J., Nuijeins, J. H., Swaak, A. J. and Hack, C. E. (1991). Predominant role of neutrophils in the inactivation of qmacroglobulin in arthritic joints. Arthritis Rheum. 34, 1139-1150.

Adams, S. I. (1988). The medical leech: a page from the anellids of internal medicine. Ann. Intern. Med. 109, 399-405.

Amiconi, G., Ascenzi, P., Bolognesi, M., Menegatti, E. and Guarneri, M. (1987). Interaction between serine (pro)enzymes and macromolecular inhibitors: thermodynamic and kinetic aspects. In: Macromolecular Biorecognition: Principles and Methods (I. M. Chaiken, E. Chiancone, A. Fontana and P. Neri, eds), pp. 117-130, The Humana Press, Clifton.

Antonini, E., Ascenzi, P., Menegatti, E. and Guarneri, M. (1983a). Multiple inter- mediates in the reaction of bovine B-trypsin with bovine pancreatic trypsin inhibitor (Kunitz). Biopolymers 22, 363-375.

Antonini, E., Ascenzi, P., Bolognesi, M., Menegatti, E. and Guarneri, M. (1983b). Transient removal of proflavine inhibition of bovine P-trypsin by the bovine basic pancreatic trypsin inhibitor (Kunitz). J. Biol. Chem. 258, 467tf-4678.

Appella, E., Weber, I. T. and Blasi, F. (1988). Structure and function of epidermal growth factor-like regions in proteins. FEBS Lett. 231, 1-4.

Ascenzi, P., Aducci, P., Amiconi, G., Ballio, A., Guaragna, A., Menegatti, E., Schnebli, H. P. and Bolognesi, M. (1991). Binding of the recombinant proteinase inhibitor eglin c from leech Hirudo medicinalis to serine (pro)enzymes: a comparative thermodynamic study. J. Mol. Recogn. 4, 113-119.

Ascenzi, P., Amiconi, G., Bolognesi, M., Menegatti, E. and Guarneri, M. (1986). Binding of porcine pancreatic secretory trypsin inhibitor to bovine P-trypsin: a kinetic study. Biopolymers 25, 2325-2333.

287


Ascenzi, P., Amiconi, G., Coletta, M., Lupidi, G., Menegatti, E., Onesti, S. and Bolognesi, M. (1992a). Binding of hirudin to human (Y, p and y-thrombin: a comparative kinetic and thermodynamic study. J. Mol. Biol. 225, 177-184.

Ascenzi, P., Fruttero, R., Amiconi, G., Pugliese, L., Bolognesi, M., Coletta, M., Onesti, S., Guarneri, M. and Menegatti, E. (1992b). Inhibition of human CX-, l3- and y-thrombin by mono-, bis-, tris- and tetra-benzamidine structures: thermodynamic study. J. Enzyme Inhibit. 6, 131-139.

Ascenzi, P., Coletta, M., Amiconi, G., De Cristofaro, R., Bolognesi, M., Guarneri, M. and Menegatti, E. (1988). Binding of the bovine basic pancreatic trypsin inhibitor (Kunitz) to human (Y-, p- and y-thrombin: a kinetic and thermodynamic study. Biochim. Biophys. Acta 956, 156-161.

Atkins, P. W. (1994). Physical Chemistry, 5th edn, Oxford University Press, Oxford.

Auerswald, E. A., Horlein, D., Reinhardt, G., Schriider, W. and Schnabel, E. (1988). Expression, isolation and characterization of recombinant (Arg-15, Glu- 52) aprotinin. Biol. Chem. Hoppe-Seyler 369(Suppl.), 27-35.

Auerswald, E. A., Morenweiser, R., Sommerhoff, C. P., Piechottka, G. P., Eckerskorn, C., Gtirtler, L. G. and Fritz, H. (1994). Recombinant leech-derived tryptase inhibitor: construction, production, protein chemical characterization and inhibition of HIV-l replication. Biol. Chem. Hoppe-Seyler 375, 695-703.

Ayala, Y. and Di Cera, E. (1994). Molecular recognition by thrombin: role of the slow + fast transition, site-specific ion binding energetics and thermodynamic mapping of structural components. J. Mol. Biol. 235, 733-746.

Badimon, L., Meyer, B. J. and Badimon, J. J. (1994). Thrombin in arterial thrombosis. Haemostasis 24, 69-80.

Bagdy, D., Barabas, E., Graf, L., Peterson, T. E. and Magnusson, S. (1976). Hirudin. Meth. Enzymol. 45, 669-678.

Baird, B. R., Cheronis, J. C., Sandhaus, R. A., Berger, E. M., White, C. W. and Repine, J. E. (1986). 0, Metabolites and neutrophil elastase synergistically cause edematous injury in isolated rat lungs. J. Appl. Physiol. 61, 2224-2229.

Banda, M., Sinha, S. and Travis, J. (1987). Inactivation of human q-proteinase inhibitor by macrophage elastase. J. Clin. Invest. 79, 1314-1318.

Banner, D., Ackermann, J., Gast, A., Gubernator, K., Hadvary, P., Hilpert, K., Labler, L., Muller, K., Schmid, G., Tschopp, T. (1994). Serine proteases: 3D structures, mechanisms of action and inhibitors. In: Perspectives in Medicinal Chemistry: Proceedings of the XZZth Znternational Symposium on Medicinal Chemistry (B. Testa, E. Kyburz, W. Fuhrer and R. Giger, eds), pp. 27-43. Chimica Acta, Basel.


Bar-Shavit, R., Eldon, A. and Vlodavski, I. (1989). Binding of thrombin to sub- endothelial extracellular matrix: protection and expression of functional properties. .Z. Clin. Invest. 84, 1096-1104.

Batchelor, A., Davidson, P. and Sully, L. (1984). The salvage of congested skin flaps by the application of leeches. Br. J. Plust. Surg. 37, 358-360.

Bates, M., Kneer, K. and Logan, C. (1989). Medicinal leech therapy: an overview. Orthop. Nursing 8, 12-16.

Beatty, K., Bieth, J. and Travis, J. (1980). Kinetics of association of serine proteinases with native and oxidized a,-proteinase inhibitor and au,-antichymotrypsin. .Z. Biol. Chem . 255, 393 l-3934.

Becker, R. C. (1992). Thrombin antagonists and antiplatelet agents. Am. 1. Cardiol. 69, 39A-50A.

Berrettini, M., Lammle, B. and Griffin, J. H. (1987). Initiation of coagulation and relationships between intrinsic and extrinsic coagulation pathways. In: Thrombosis and Huemostasis (M. Verstraete, J. Vermylen, R. Lijnen and J. Arnout, eds), pp. 473-495. Leuven University Press.

Betz, A., Hofsteenge, J. and Stone, S. R. (1991a). Ionic interactions in the formation of the thrombin-hirudin complex. Biochem. .I. 275, 801-803.

Betz, A., Hofsteenge, J. and Stone, S. R. (1991b). Role of interactions involving C-terminal nonpolar residues of hirudin in the formation of the thrombin-hirudin complex. Biochemistry 30, 9848-9853.

Betz, A., Hofsteenge, J. and Stone, S. R. (1992). pH Dependence of the interaction of hirudin with thrombin. Biochemistry 31, 1168-1172.

Betz, A., Hopkins, P. C., Le Bonniec, B. F. and Stone, S. R. (1994). Contribution of interactions with the core domain of hirudin to the stability of its complex with thrombin. Biochem. .I. 298, 507-510.

Betzel, C., Dauter, Z., Genov, N., Lamzin, V., Navaza, J., Schnebli, H. P., Visanji, M. and Wilson, K. S. (1993). Structure of the proteinase inhibitor eglin c with hydrolysed reactive centre at 2.0 8, resolution. FEBS Lett. 317, 185-188.

Bichler, J., Gemmerli, R. and Fritz, H. (1991). Studies for revealing a possible sensitization to hirudin after repeated intravenous injection in baboons. Thromb. Res. 61, 39-51.

Bieth, J. (1974). Some kinetic consequences of the tight binding of protein:proteinase- inhibitors to proteolytic enzymes and their application to the determination of dissociation constants. In: Buyer Symposium V on Proteinase Inhibitors (H. Fritz, H. Tschesche, L. J. Greene and E. Truscheit, eds), pp. 463-469. Springer, Berlin.


Bieth, J. (1980). In vivo significance of kinetic constants of macromolecular proteinase inhibitors. Bull. Eur. Physiopath. Resp. 16, 183-195.

Bigler, T. L., Lu, W., Park, S. J., Tashiro, M., Wieczorek, M., Wynn, R. and Laskowski, M. Jr (1993). Binding of amino acid side chains to preformed cavities: interaction of serine proteinases with turkey ovomucoid third domains with coded and noncoded P, residues. Protein Sci. 2, 786-799.

Bode, W. and Huber, R. (1991). Ligand binding: proteinase-protein inhibitor interactions. Curr. Opin. Struct. Biol. 1, 45-52.

Bode, W. and Huber, R. (1992). Natural protein proteinase inhibitors and their interaction with proteinases. Eur. J. Biochem. 204, 433-451.

Bode, W., Huber, R. Rydel, T. J. and Tulinsky, A. (1992a). X-Ray crystal structures of human a-thrombin and of human thrombin-hirudin complex. In: Thrombin: Structure and Function (L. J. Berliner, ed.), pp. 3-61. Plenum Press, New York.

Bode, W., Turk, D. and Karshikov, A. (1992b). The refined 1.9A X-ray crystal structure of D-Phe-Pro-Arg chloromethylketone inhibited human a-thrombin: structure analysis, overall structure, electrostatic properties, detailed active site geometry, structure-function relationships. Protein Sci. 1, 426-471.

Bode, W., Mayr, I., Baumann, U., Huber, R., Stone, S. R. and Hofsteenge, J. (1989). The refined 1.9 A crystal structure of human a-thrombin. Interaction with D-Phe-Pro-Arg chlorome-thylketone and significance of the tyr-pro-pro-trp insertion segment. EMBO J. 8, 3467-3475.

Bode, W., Papamokos, E., Musil, D., Seemuller, U. and Fritz, H. (1986). Refined 1.2 8, crystal structure of the complex formed between subtilisin Carlsberg and the inhibitor eglin c: molecular structure of eglin and its detailed interaction with subtilisin. EMBO J. 5, 813-818.

Bode, W., Papamokos, E. and Musil, D. (1987). The high-resolution X-ray crystal structure of the complex formed between subtilisin Carlsberg and eglin c, an elastase inhibitor from the leech Hirudo medicinalis: structural analysis, subtilisin structure and interface geometry. Eur. J. Biochem. 166, 673-692.

Bodong, A. (1905) ijber Hirushin. Arch. Exp. Pathol. Pharmacol. 52, 242-261.

Bolognesi, M., Ascenzi, P., Amiconi, G., Menegatti, E. and Guarneri, M. (1987). Structural bases for the recognition of inhibitors by serine proteinases and their zymogens. In: Macromolecular Biorecognition: Principles and Methods (I. M. Chaiken, E. Chiancone, A. Fontana and P. Neri, eds), pp. 81-100. The Humana Press. Clifton.

Bolognesi, M., Gatti, G., Menegatti, E., Guarneri, M., Marquart, M., Papamokos, E. and Huber, R. (1982). Three-dimensional structure of the complex between


pancreatic secretory inhibitor (Kazal-type) and trypsinogen at 1.8 8, resolution. Structure solution, crystallographic refinement and preliminary structural interpretation. 1. Mol. Biol. 162, 839-868.

Bolognesi, M., Pugliese, L., Gatti, G., Frigerio, F., Coda, A., Antolini, L., Schnebli, H. P., Menegatti, E., Amiconi, G. and Ascenzi, P. (1990). X-ray crystal strusture of the bovine a-chymotrypsin/eglin c complex at 2.6 A resolution. 1. Mol. Recogn. 3, 163-168.

Braun, N. J. and Schnebli, H. P. (1986). A brief review of the biochemistry and pharmacology of eglin c, an elastase inhibitor. Eur. J. Respir. Dis. 146(Suppl.), 541-547.

Braun, N. J. and Schnebli, H. P. (1987). Interaction of eglin c with polymorphonuclear cells: evidence for binding to the cell surface. Biol. Chem. Hoppe-Seyler 368, 155-161.

Braun, N. J., Bodmer, J. L., Virca, G. D., Metz-Virca, G., Maschler, R., Bieth, G. and Schnebli, H. P. (1987). Kinetic studies on the interaction of eglin c with human leukocyte elastase and cathepsin G. Biol. Chem. Hoppe-Seyler 368, 299-308.

Braun, P. J., Dennis, S., Hofsteenge, J. and Stone, S. R. (1988). Use of-site-directed mutagenesis to investigate the basis for the specificity of hirudin. Biochemistry 27, 6517-6522.

Butkowski, R. J., Elion, J., Downing, M. R. and Mann, K. G. (1977). Primary structure of prethrombin 2 and ol-thrombin. J. Biol. Chem. 252, 4942-4957.

Cannon, P. C., McCabe, C. H., Henry, T. D., Rogers, W. J., Schweiger, M., Gibson, R. S., Anderson, J. L., Williams, D. 0. and Braunwald, E. (1993). Thrombolysis in acute myocardial infarction. J. Am. Coll. Cardiol. 21, 740-741.

Chang, J. Y. (1983). The functional domain of hirudin: a thrombin-specific inhibitor. FEBS Lett. 164, 307-313.

Cheronis, J. C. and Repine, J. E. (eds) (1993). Proteases, Protease Znhibitors and Protease-Derived Peptides, Birkhauser, Basel.

Chesebro, J. H. and Fuster, V. (1987). The therapeutic challenge of plaque rupture: value of biochemical markers. J. Am. Coil. Curdiof. 10, 1005.

Chignard, M., Balloy, V. and Renesto, P. (1993). Leukocyte elastase-mediated release of von Willebrand factor from cultured endothelial cells. Eur. Respir. J. 6, 791-796.

Church, F. C., Phillips, J. E. and Woods, J. L. (1991). Chimeric antithrombin peptide: characterization of an Arg-Gly-Asp- and hirudin carboxyl terminus-containing synthetic peptide. J. Biol. Chem. 266, 11975-11979.


Clore, G. M., Gronenborn, A. M., Kjaer, M. and Poulsen, F. M. (1987a). The determination of the three-dimensional structure of barley serine proteinase inhibitor 2 by nuclear magnetic resonance, distance geometry and restrained molecular dynamics. Protein Eng. 1, 305-311.

Clore, G. M., Gronenborn, A. M., James, M. N. G., Kjaer, M., McPhalen, C. A. and Poulsen, F. M. (1987b). Comparison of the solution and X-ray structures of barley serine proteinase inhibitor 2. Protein. Eng. 1, 313-318.

Coletta, M., Ascenzi, P., Amiconi, G., Bolognesi, M., Guarneri, M. and Menegatti, E. (1990). Bovine trypsinogen activation; a thermodynamic study. Biophys. Chem. 37, 355-362.

Collet, D., Lefebvre, F., Quentin, C. and Rabaud, M. (1991). In vitro studies of elastin-fibrin biomaterial degradation: preservative effects of protease inhibitors and antibiotics. Biomaterials 12, 763-766.

Connolly, M. L. (1992). Shape distribution of protein topography. Biopolymers 32, 1215-1236.

Creighton, T. E. and Darby, N. J. (1989). Functional evolutionary divergence of proteolytic enzymes and their inhibitors. TZBS 14, 319-324.

Cutruzzola’, F., Ascenzi, P., Barra, D., Bolognesi, M., Menegatti, E., Sarti, P., Schnebli, H. P., Tomova, S. and Amiconi, G. (1993). Selective oxidation of Met-192 in bovine a-chymotrypsin: effect on catalytic and inhibitor binding properties. Biochim. Biophys. Acta 1161, 201-208.

Dautzer, Z., Betzel, Ch., Hohne, W. E., Ingelman, M. and Wilson, K. S. (1988). Crystal Structure of a complex between thermitase from Thermoactinomyces vulgaris and the leech inhibitor eglin c. FEBS Lett. 236, 171-178.

Dauter, Z., Betzel, Ch., Genov, N., Pipon, N. and Wilson, K. S. (1991). Complex between the subtilisin from a mesophilic bacterium and the leech inhibitor eglin c. Acta Crystallogr. B47, 707-730.

Davie, E. W., Fujikawa, K. and Kisiel, W. (1991). The coagulation cascade: inhibition, maintenance and regulation. Biochemistry 30, 10364-10369.

De Cristofaro, R., Fenton, J. W. II and Di Cera, E. (1992). Modulation of thrombin-hirudin interaction by specific ion effects. J. Mol. Biol. 226, 263-269.

De Cristofaro, R., Rocca, B., Bizzi, B. and Landolfi, R. (1993). The linkage between binding of the C-terminal domain of hirudin and amidase activity in human a-thrombin. Biochem. J. 289, 485-490.

Dean, R. T. and Schnebli, H. P. (1989). Control of exogenous proteinases and their inhibitors at the macrophage cell surface. Biochim. Biophys. Acta 992, 174-180.


Dean, R. T., Nick, H. P. and Schnebli, H. P. (1989). Free radicals inactivate human neutrophil elastase and its inhibitors with a comparable efficiency. Biochem. Biophys. Rex Commun. 159, 821-827.

Degen, S. J. F., Gillivray, R. T. A. and Davie, D. W. (1983). Characterization of the complementary deoxyribonucleic acid and gene coding for human prothrombin. Biochemistry 22, 2087-2097.

Dennis, S., Wallace, A., Hofsteenge, J. and Stone, S. R. (1990). Use of fragments of hirudin to investigate thrombin-hirudin interaction. Eur. J. Biochem. 188, 61-66.

Devereux, J., Haeberli, P. and Smithies, 0. (1984). A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12, 387-395.

Dodt, J., Koler, S. and Baici, A. (1988). Interaction of site specific hirudin variants with ol-thrombin. FEBS Lett. 229, 87-90.

Dodt, J., Kohler, S., Schmitz, T. and Wilhelm, B. (1990). Distinct binding sites of ala48-hirudini-17 and alads-hirudin48-ss on a-thrombin. J. Biol. Chem. 265, 713-718.

Dodt, J., Miiller, H., Seemtiller, U. and Chang, J.-Y. (1984). The complete amino acid sequence of hirudin: a thrombin specific inhibitor. FEBS Lett. 165, 180-183.

Dodt, J., Seemiiller, U., Maschler, R. and Fritz, H. (1985). The complete covalent structure of hirudin: localization of the disulfide bonds. Biol. Chem. Hoppe-Seyler 366, 379-385.

Dodt, J., Seemfiller, U. and Fritz, H. (1987). Influence of chain shortening on the inhibitor properties of hirudin and eglin c. Biol. Chem. Hoppe-Seyler 368, 1447-1453.

Dunwiddie, C. T., Waxman, L., Vlasuk, G. P. and Friedman, P. A. (1993). Purification and characterization of inhibitors of blood coagulation factor Xa from haematophagus organisms. Meth. Enzymol. 223, 291-312.

Duswald, K. H., Jochum, M. and Fritz, H. (1982). Pathophysiology and pathobiochemistry of sepsis and septic shock: new findings on the importance of proteinase release and the possibilities for treatment. Intern. Devel. Diagn. Therapy 2, 1-8.

Eklund, K. K. and Stevens, R. L. (1993). Mast cell proteases. In: innovations in Proteases and Their Inhibitors (F. X. Avils, ed.), pp. 241-257. Walter de Gruyter, Berlin.

Electricwala, A., Hartwell, R., Scawen, M. D. and Atkinson, T. (1993a). The complete amino acid sequence of a hirudin variant from the leech Hirudinaria manillensis. J. Protein Chem. 12, 365-370.


Electricwala, A., von Sicard, N. A. E., Sawyer, R. T. and Atkinson, T. (1993b). Biochemical characterization of a pancreatic elastase inhibitor from the leech Hirudinaria manillensis. 1. Enzyme Inhib. 6, 293-302.

Elion, J., Boissel, J.-P., Le Bonniec, B., Bezeaud, A., Jondrot-Petrus, M., Rabiet, M.-J. and Guillin, M.-C. (1986). Proteolytic derivatives of thrombin. Ann. N. Y. Acad. Sci. 485, 16-26.

Engemann, J. F. and Hegner, R. W. (1981). Phylum annelida, class III: Hirudo medicinalis, the medical leech. In: Invertebrate Zoology (3rd edn), pp. 420-426. MacMillan, New York.

Esmon, N. L., Owen, W. G. and Esmon, C. T. (1982). Isolation of a membrane-bound cofactor for thrombin-catalyzed activation of protein C. J. Biol. Chem. 257, 859-864.

Esposito, A. L., Clark, C. A., Poirier, W. J. and Kephart, P. A. (1988). Protease inhibitor eglin c affects superoxide anion release but not bacterial killing by human polymorphonuclear leukocytes. Exp. Lung Res. 14, 743-756.

Esposito, A. L., Quinn, L. A., Lucey, E. C. and Snider, G. L. (1987). Effect of the elastase inhibitor, eglin c, on antibacterial mechanisms in experimental pneumonia: description of a system to quantitate phagocytic and bactericidal activity of resident murine alveolar macrophages in vivo. Am. Rev. Respir. Dis. 135, 676-681.

Faller, B., Dirrig, S., Rabaud, M. and Bieth, J. G. (1990). Kinetics of the inhibition of human pancreatic elastase by recombinant eglin c. Biochem. .I. 270, 639-644.

Fareed, J., Walenga, J. M., Hoppensteadt, D. and Pifarre, R. (1989). Development perspectives for recombinant hirudin as an antithrombotic agent. Biol. Clin. Hematol. 11, 143-152.

Fenton, J. W. II (1986). Thrombin. Ann. N. Y. Acad. Sci. 485, 5-15.

Fenton, J. W. II (1988). Thrombin bioregulatory functions. Adv. Cfin. Enzymof. 6, 186-193.

Fenton, J. W. II (1989). Thrombin interactions with hirudin. Semin. Thromb. Hemost. 15, 265-268.

Fink, E., Nettelbeck, R. and Fritz, H. (1986a). Inhibition of mast cell chymase by eglin c and antileukoprotease (HUSI-I). Biol. Chem. Hoppe-Seyler 367, 567-571.

Fink, E., Rehm, H., Gippner, C., Bode, W., Eulitz, M., Machleidt, W. and Fritz, H. (1986b). The primary structure of bdellin B-3 from the leech Hirudo medicinalis: bdellin B-3 is a compact proteinase inhibitor of a ‘non-classical’ Kazal type; it is present in the leech in a high molecular mass form. Biol. Chem. Hoppe-Seyler 367, 1235-1242.

Proteinase Inhibitors from Leech Hirudo mecficinaiis 295

Finkelstein, A. V. and Janin, J. (1989). The price of lost freedom: entropy of bimolecular complex formation. Protein Eng. 3, l-3.

Fiorucci, L., Erba, F., Coletta, M. and Ascoli, F. (1995). Evidence for multiple interacting binding sites in bovine tryptase. FEBS Lett. 363, 81-84.

Folkers, P. J. M., Clore, G. M., Driscoll, P. C., Dodt, J., Kohler, S. and Gronenborn, A. M. (1989). Solution structure of recombinant hirudin and the Lys47 - > Glu mutant. A nuclear magnetic resonance and hybrid distance geometry dynamical simulated annealing study. Biochemistry 28, 2601-2617.

Foucher, G., Henderson, H. R., Mancau, M., Merle, M. and Braun, F. M. (1981). Distal digital replantation: one of the best indications for microsurgery. Znt. J. Microsurg. 3, 263-270.

Fox, I., Dawson, A., Lovnds, P., Eisner, J., Findlen, K., Levin, E., Hanson, D., Mant, T., Wagner, J. and Maraganore, J. (1993). Anticoagulant activity of hirulog, a direct thrombin inhibitor, in humans. Thromb. Huemostus. 69, 157-163.

Friedrich, T., Kroger, B., Biolajan, S., Lemaire, H. G., Hoffken, H. W., Reuschenbach, P., Otte, M. and Dodt, J. (1993). A Kazal-type inhibitor with thrombin specificity from Rhodnius prolixus. J. Biol. Chem. 268, 16216-16222.

Frigerio, F., Coda, A., Pugliese, L., Lionetti, C., Menegatti, E., Amiconi, G., Schnebli, H. P., Ascenzi, P. and Bolognesi, M. (1992). Crystal and molecular structure of the bovine a-chymotrypsin-eglin c complex at 2.0 8, resolution. J. Mol. Biol. 225, 107-123.

Fritz, H. (1988). Human mucus proteinase inhibitor. Biol. Chem. Hoppe-Seyler 369 (Suppl.), 79-82.

Fritz, H., Oppitz, K. H., Gebhardt, M., Oppitz, I., Werle, E. and Marx, R. (1969). ijber das vorkommen eines trypsin-plasmin-inhibitors in hirudin. Hoppe-Seyler’s 2. Physiol. Chem. 350, 91-92.

Frommherz, K. J., Faller, B., and Bieth, J. G. (1991). Heparin strongly decreases the rate of inhibition of neutrophil elastase by a,-proteinase inhibitor. J. Biol. Chem. 266, 1535615362.

Fujinaga, M., Sielecki, A., Read, R. J., Ardelt, W., Laskowski, M. Jr and James, M. N. G. (1987). Crystal and molecular structures of the complex of a-chymotrypsin with its inhibitor turkey ovomucoid third domain at 1.8 A resolution. J. Mol. Biol. 195, 397-418.

Fukusen, N., Kato, J., Kido, H. and Katunuma, N. (1987). Kinetic studies on the inhibition of mast cell chymase by natural serine protease inhibitors: indications for potential biological functions of these inhibitors. Biochem. Med. Metub. Biol. 38, 165-169.


Gardell, S. J. and Friedman, P. A. (1993). Vampire bat salivary plasminogen activator. Meth. Enzymol. 223, 233-249.

Garrison, F. M. (1914). An Zntroduction to the History of Medicine. Saunders Co., Philadelphia.

Goddard, P., O’Mullane, J., Ambler, L., Daw, A., Brookman, L., Lee, A. and Petrak, K. (1991). R-[N-acetylleglin c:poly(oxyethylene) conjugates: preparation, plasma persistence, and urinary excretion. J. Pharm. Sci. 80, 1171-1176.

Goldstein, A. M., Murer, E. H. and Weinbaum, G. (1986). Serine protease inhibitors of north american leechs. Comp. Biochem. Physiol. 84B, 117-124.

Gros, P., Fujinaga, M., Dijkstra, B. W., Kalk, K. H. and Hol, W. G. (1989a). Crystallographic refinement by incorporation of molecular dynamics: thermostable serine proteinase thermitase complexed with eglin c. Acta Crystallogr. B45, 488-499.

Gros, P., Betzel, C., Dauter, Z., Wilson, K. S. and Hol, W. G. (1989b). Molecular dynamics refinement of a thermitase-eglin c complex at 1.98 A resolution and comparison of two crystal forms that differ in calcium content. .I. Mol. Biol. 210, 347-367.

Gros, P., Teplyakov, A. V. and Hol, W. G. (1992). Effects of eglin c binding to thermitase: three-dimensional structure comparison of native thermitase and thermitase eglin c complexes. Proteins 12, 63-74.

Griitter, M. G., Priestle, J. P., Rahuel, J., Grossenbacher, H., Bode, W., Hofsteenge, J. and Stone, S. R. (1990). Crystal structure of the thrombin-hirudin complex. A novel mode of serine proteinase inhibition. EMBO .I. 9, 2361-2365.

Hansen, P. R. and Stawski, G. (1994). Neutrophil mediated damage to isolated myocites after anoxia and reoxygenation. Cardiovasc. Res. 28, 565-569.

Harmon, J. T. and Jamieson, G. A. (1985). Thrombin binds to a high-affinity 900000 dalton site on human platelets. Biochemistry 24, 58-64.

Haruyama, H. and Wiithrich, K. (1989). The conformation of recombinant desulfatohirudin in aqueous solution determined by nuclear magnetic resonance. Biochemistry 28, 43014312.

Harvey, R. P., Degryse, E., Stefani, L., Schamber, F., Cazenave, J. P., Courtney, M., Tolstoshev, P. and Lecocq, J. P. (1986). Cloning and expression of a cDNA coding for the anticoagulant hirudin from the bloodsucking leech, Hirudo medicinalis. Proc. Natl. Acad. Sci. U.S.A. 83, 1084-1088.

Hauptmann, J. and Kaiser, B. (1993a). Anticoagulant and antithrombotic action of the factor Xa inhibitor antistasin (ATS). Thromb. Res. ‘71, 169-174.

Proteinase Inhibitors from Leech Hitudo medicinalis 297

Hauptmann, J. and Kaiser, B. (1993b). Anticoagulant potential of synthetic and recombinant inhibitors of factor Xa and thrombin in vitro. Blood Cougul. Fibrinolysis 4, 577-582.

Haycraft, J. B. (1884). ijber die einwirkung eines secretes des officinalen blutegels auf die gerinnbarkeit des blutes. Naunyn-Schmiedebergs Arch. Exp. Pathol. Pharmakol. 18, 209-217.

Hecht, H. J., Szardenings, M., Collins, J. and Schomburg, D. (1991). Three- dimensional structure of the complexes between bovine chymotrypsinogen A and two recombinant variants of human pancreatic secretory trypsin inhibitor (Kazal-type). J. Mol. Biol. 220, 711-722.

Heidtmann, H. and Travis, J. (1986). Human q-proteinase inhibitor. In: Proteiruzse Inhibitors (A. J. Barrett and G. Salvesen, eds), pp. 442-456. Elsevier, Amsterdam, New York. Oxford.

Heinz, D. W., Hyberts, S. G., Peng, J. W., Priestle, J. P., Wagner, G. and Griitter, M. G. (1992). Changing the inhibitory specificity and function of the proteinase inhibitor eglin c by site-directed mutagenesis: functional and structural investigation. Biochemistry 31, 8755-8766.

Heinz, D. W., Liersh, M. and Griitter, M. G. (1989). Crystallization of human leukocyte elastase with its inhibitor Pro44-eglin c. J. Mol. Biol. 207, 641-642.

Heinz, D. W., Priestle, J. P., Rahuel, J., Wilson, K. S. and Grutter, M. G. (1991). Refined crystal structures of subtilisin Novo in complex with wild-type and two mutant eglins: comparison with other serine proteinase complexes. J. Mol. Biol. 217, 353-371.

Hipler, K., Priestle, J. P., Rahuel, J. and Grtitter, M. G. (1992). X-ray crystal structure of the serine proteinase inhibitor eglin c at 1.95 A resolution. FEBS Lett. 309, 139-145.

Hirsh, J. (1991). Heparin. N. Engl. J. Med. 324, 1565-1574.

Horn, H. and Heidland, A. (eds) (1982). Proteases: Potential Role in Health and Diseases. Plenum Press, New York, London.

Ho, C. K. and Fersht, A. R. (1986). Internal thermodynamics of position 51 mutants and natural variants of tyrosyl-tRNA synthetase. Biochemistry 25, 1891-1897.

Hoffmann, A. and Markwardt, F. (1984). Inhibition of the thrombin-platelet reaction by hirudin. Haemostasis 14, 164-169.

Hofmann, K. J., Nutt, E. M. and Dunwiddie, C. T. (1992). Site-directed mutagenesis of the leech-derived factor Xa inhibitor antistasin. Biochem. J. 287, 943-949.


Huber, R. and Bode, W. (1978). Structural basis of activation and action of trypsin. Act. Chem. Res. 11, 114-122.

Hyberts, S. G. and Wagner, G. (1990). Sequence-specific 1H NMR assignments and secondary structure of eglin c. Biochemistry 29, 1465-1474.

Hyberts, S. G., Goldberg, M. S., Havel, T. F. and Wagner, G. (1992). The solution structure of eglin c based on measurements of many NOE and coupling constants and its comparison with X-ray structures. Protein Sci. 1, 736-751.

Hyberts, S. G., Marki, W. and Wagner, G. (1987). Stereospecific assignments of side-chain protons and characterization of torsion angles in eglin c. Eur. J. Biochem. 164, 625-635.

Inauen, W., Granger, D. H., Meininger, C. J., Schelling, M. E., Granger, H. J. and Evietys, P. R. (1990). Anoxia-reoxygenation-induced, neutrophyl-mediated endothelial cell injury: role of elastase. Am. J. Physiol. 259, H925-H931.

Jackman, M. P., Parry, M. A. A., Hofsteenge, J. and Stone, S. R. (1992). Intrinsic fluorescence changes and rapid kinetics of the reaction of thrombin with hirudin. J. Biol. Chem. 267, 15375-15383.

Janin, J. and Chothia, C. (1990). The structure of protein-protein recognition sites. J. Biol. Chem. 265, 16027-16030.

Janin, J. (1995). Elusive affinities. Proteins 21, 30-39.

Janoff, A. (1985a). Elastases and emphysema: current assessment of the protease- antiprotease hypothesis. Am. Rev. Resp. Dis. 132, 417-434.

Janoff, A. (1985b). Elastase in injury. Annu. Rev. Med. 36, 207-216.

Janus, T. J., Lewis, S. D., Lorand, L.and Shafer, J. A. (1983). Promotion of thrombin- catalyzed activation of factor XIII by fibrinogen. Biochemistry 22, 6269-6272.

Jochum, M., Witte, J., Schiessler, H., Selbmann, H. K., Ruckdeschl, G. and Fritz, H. (1981). Clotting and other plasma factors in experimental endotoxemia: inhibition of degradation by exogenous proteinase inhibitors. Eur. Surg. Res. 13, 152-168.

Jochum, H., White, J., Duswald, K. H., Inthorn, D., Welter, H. and Fritz, H. (1986). Pathobiochemistry of sepsis: role of proteinases, proteinase inhibitors and oxidizing agents. Behring Inst. Mitt. Feb, 121-130

Jonassen, I. and Svendsen, I. (1982). Identification of the reactive sites in two homologous serine proteinase inhibitors isolated from barley. Curlsberg Res. Commun . 47, 199-203.


Johnson, P. H. (1994). Hirudin: clinical potential of a thrombin inhibitor. Annu. Rev. Med. 45, 165-177.

Johnson, D. and Travis, J. (1978). Structural evidence for methionine at the reactive site of human or-proteinase inhibitor. J. Biol. Chem. 253, 7142-7145.

Johnson, D. and Travis, J. (1979). The oxidative inactivation of human ol,-proteinase inhibitor. J. Biol. Chem.254, 4022-4026.

Junger, W. G., Hallstrom, Redl, H. and Schlag, G. (1992). Inhibition of human, ovine, and baboon neutrophil elastase with eglin c and secretory leukocyte proteinase inhibitor. Biol. Chem. ffoppe-Seyler 373, 119-122.

Junger, W. G., Lieners, C., Redl, H. and Schlag, G. (1989). Reasons for the ineffectiveness of eglin c to ameliorate endotoxin shock in sheep. Prog. Clin. Biol. Res. 308, 953-957.

Kanayama, N. and Terao, T. (1990). Inactivation of human tumour cell pro-urokinase by granulocyte elastase. Jpn J. Cancer Res. 81, 994-1002.

Kao, R. C., Wehner, N. G., Skubitz, K. M., Gray, B. H. and Hoidal, J. R. (1988). Proteinase 3: a distinct human polymorphonuclear leukocyte proteinase that produces emphysema in hamsters. J. Clin. Invest. 82, 1963-1973.

Karlinsky J. B. and Snider, G. L. (1978). Animal models of emphysema. Am. Rev. Respir. Dis. 117, 1099-1108.

Karshikov, A., Bode, W., Tulinsky, A. and Stone, R. (1992). Electrostatic interactions in the association of proteins: an analysis of the thrombin-hirudin complex. Protein Sci. 1, 727-735.

Kawabata, S., Morita, T., Iwanaga, S. and Igarashi, H. (1985). Staphylocoagulase- binding region in human prothrombin. J. Biochem. 97, 325-331.

Kawaji, K., Matsunaga, Y., Kido, H. and Ogura, T. (1993). Inhibition of eosinophil peroxidase release by chymotrypsin type protease.inhibitors from activated human eosinophils. Arerugi 42, 1591-1599.

Kelly, A. B., Marzec, U. M., Krupski, W., Bass, A., Cadroy, Y., Hanson, S. R. and Harker, L. A. (1991). Hirudin interruption of heparin-resistant arterial thrombus formation in baboons. Blood 77, 10061012.

Kline, T., Hammond, L., Bourdon, P. and Maraganore, J. M. (1991). Hirulog peptides with scissile bond replacements resistant to thrombin cleavage. Biochem. Biophys. Res. Commun. 177, 1049-1055.

Knapp, A., Degenhardt, T. and Dodt, J. (1992). Hirudisins: hirudin-derived thrombin inhibitors with disintegrin activity. J. Biol. Chem. 267, 24230-24234.


Krezei, A. M., Wagner, G., Seymour-Ulmer, J. and Lazarus, R. A. (1994). Structure of the RGD protein decorsin: conserved motif and dinstinct function in leech proteins that affect blood clotting. Science 264, 1944-1947.

Krstenansky, J. L. and Mao, S. J. (1987). Antithrombin properties of C-terminus of hirudin using synthetic unsulfated N-a-acetyl hirudin 45-65. FEBS Lett. 211, 10-16.

Krstenansky, J. L., Owen, T. J., Yates, M. T. and Mao, S. J. (1988a). Comparison of hirudin and hirudin PA C-terminal fragments and related analogs as antithrombin agents. Thromb. Res. 52, 137-141.

Krstenansky, J. L., Owen, T. J., Yates, M. T. and Mao, S. J. T. (1988b). Design, synthesis and antithrombin activity for conformationally restricted analogs of peptide anticoagulants based on the C-terminal region of the leech peptide, hirudin. Biochim. Biophys. Acta 957, 53-59.

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, 661-670.

Lai, Y. L. and Diamond, L. (1990). Inhibition of porcine pancreatic elastase-induced emphysema by eglin c. Exp. Lung. Res. 16, 547-557.

Laskowski, M. Jr and Kato, I. (1980). Protein inhibitors of proteinases. Annu. Rev. Biochem . 49, 593-626.

Lau, P. P., Van Handel, M., Larvin, M., McMahon, M. J. and Geokas, M. C. (1990). Proteolytic degradation of human recombinant proinsulinhnsulin by sera from acute pancreatitis patients and complete inhibition by eglin c. Pancreas 5, 17-26, 240.

Lazar, J. B., Winant, R. C. and Johnson, P. H. (1991). Hirudin: amino-terminal residues play a major role in the interaction with thrombin. J. Biol. Chem. 266, 685-688.

Lid6n, R. M., ThCroux, P., Juneau, M., Adelman, B. and Maraganore, J. M. (1993). Initial experience with a direct antithrombin, hirulog, in unstable angina: anticoagulant, antithrombotic, and clinical effects. Circulation 88, 1495-1501.

Lim, I. (1986). Successful transfer of ‘free’ microvascular superficial temporal artery flap with no obvious venous drainage and use of leeches for reducing venous congestion: case report. Microsurgery 7, 87-88.

Linas, S. L., Whittenburg, D., Parsons, D. E. and Repine, J. E. (1992). Mild renal ischemia activates primed neutrophils to cause acute renal failure. Kidney Znt. 42, 610-616.


Linas, S. L., Whittenburg, D. and Repine, J. E. (1991). Role of neutrophil derived oxidants and elastase in lipopolysaccharide-mediated renal injury. Kidney Znt. 39, 618-623.

Loukusa, A. K., Veijola, M. and Rajaniemi, H. (1990). Plasminogen activator is involved in the hCG-induced neutrophil extravasation and vasopermeability increase in the rat testis. ht. 1. Androl. 13, 306314.

Lucey, E. C., Stone, P. J., Christensen, T. G., Breuer, R., Calore, J. D. and Snider, G. L. (1986). Effect of varying the time interval between intratracheal administration of eglin c and human neutrophil elastase on prevention of emphysema and secretory cell metaplasia in hamsters: with observations on the fate of eglin c and the effect of repeated instillations. Am. Rev. Respir. Dis. 134, 471-475.

Maier, K., Muller, H., Tesch, R., Trolp, R., Witt, I. and Holzer, H. (1979). Primary structure of yeast proteinase B inhibitor 2. 1. Biol. Chem. 254, 12555-12561.

Major, R. H. (1954). A History of Medicine. (C. C. Thomas, ed.), Springfield.

Mao, S. J., Yates, M. T., Blankenship, D. T., Cardin, A. D., Krstenansky, J. L., Lovenberg, W. and Jackson, R. L. (1987). Rapid purification and revised amino- terminal sequence of hirudin: a specific thrombin inhibitor of the bloodsucking leech. Anal. Biochem. 161, 214-218.

Marki, W. E., Grossenbacher, H., Griitter, M. G., Liersch, M. H., Meyhack, B. and Heim, J. (1991). Recombinant hirudin: genetic engineering and structure analysis. Semin. Thromb. Hemost. 17, 88-93.

Markwardt, F. (1955). Untersuchungen uber hirudin. Nuturwissenschuften 42, 537-538.

Markwardt, F. (1956). Untersuchungen uber den mechanismus der blutgerinnung- shemmenden wirkung des hirudins. Arch. Exp. Phatol. Pharmakol. 229,389-399.

Markwardt, F. (1957). Die isolierung und chemische characterisierung des hirudins. Hoppe-Seyler’s 2. Physiol. Chem. 308, 147-156.

Markwardt, F. (1959). Ueber den einflup des hirudins auf die thrombenbildung. Arch. Exp. Pathol. 242, 343-345.

Markwardt, F. (1970). Hirudin as an inhibitor of thrombin. Meth. Enzymol. 19, 924-932.

Markwardt, F. (1985). Pharmacology of hirudin: one hundred years after the first report of the anticoagulant agent in medicinal leeches. Biomed. Biochim. Actu 44, 1007-1013.

Markwardt, F. (1986). Pharmacological approaches to thrombin regulation. Ann. N. Y. Acad. Sci. 485, 204-214.


Markwardt, F. (1988). The comeback of hirudin: an old-established anticoagulant agent. Folia Haematol. Leipz 115, 10-23.

Markwardt, F. (1992). Hirudin: the promising antithrombotic. Cardiovasc. Drug Rev. 10, 21 l-232.

Markwardt, F. and Landmann, H. (1971). Blutgerinnungshemmende proteine, peptide und aminos%urederivative. Handbook Exp. Pharmacol. 27, 76-142.

Markwardt, F., Fink, E., Kaiser, B., Klocking, H. P., Nowak, G., Richter, M. and Sturzebecker, J. (1988). Pharmacological survey of recombinant hirudin. Pharmazie 43, 202-207.

Markwardt, F., Glusa, E. and Sturzebecher, J. (1990). Antithrombin effects of native and recombinant hirudins. Biomed. Biochim. Acta 49, 399-404.

Markwardt, F., Hauptmann, J., Nowak, G., Klessen, C. and Walsmann, P. (1982). Pharmacological studies on the antithrombotic action of hirudin in experimental animals. Thromb. Haemost. 47, 226-229.

Markwardt, F., Nowak, G. and Sturzebecher, J. (1991a). Clinical pharmacology of recombinant hirudin. Haemostasis 21, 133-136.

Markwardt, F., Nowak, G. and Bucha, E. (1991b). Hirudin as anticoagulant in experimental hemodialysis. Haemostasis 21, 149-155.

Marquart, M., Walter, J., Deisenhofer, J., Bode, W. and Huber, R. (1983). The geometry of the reactive site and of the peptide groups in trypsin, trypsinogen and its complexes with inhibitors. Acta Crystaffogr. B39, 480-490.

Maugeri, N., Evangelista, V., Piccardoni, P., Dell’Elba, G., Celardo, A., de Gaetano, G. and Cerletti, C. (1992). Transcellular metabolism of arachidonic acid: increased platelet thromboxane generation in the presence of activated polymorphonuclear leukocytes. Blood 80, 44745 1.

Maywald, F., Boldicke, T., Gross, G., Frank, R., Blocker, H., Meierhans, A., Schwellnus, K., Ebbers, J., Bruns, W., Reinhardt, G., Schnabel, E., Schroder, W., Fritz, H. and Collins, J. (1988). Human pancreatic secretory trypsin inhibitor (PSTI) produced in active form and secreted from Escherichia cofi. Gene 68, 357-369.

McPhalen, C. A. and James, M. N. G. (1987). Crystal and molecular structure of the serine proteinase inhibitor CI-2 from barley seeds. Biochemistry 26, 261-269.

McPhalen, C. A. and James, M. N. G. (1988). Structural comparison of two serine proteinase-protein inhibitor complexes: eglin c-subtilisin Carlsberg and ci- 2-subtilisin Novo. Biochemistry 27, 6582-6598.

Proteinase Inhibitors from Leech Himdo medicine/is 303

McPhalen, C. A., Schnebli, H. P. and James, M. N. G. (1985a). Crystal and molecular structure of the inhibitor eglin from leeches in complex with subtilisin Carlsberg. FEBS Lett. 188,’ 55-58.

McPhalen, C. A., Svendsen, I., Jonassen, I and James, M. N. G. (1985b). Crystal and molecular structure of chymotrypsin inhibitor 2 from barley seeds in complex with subtilisin Novo. Proc. Natl. Acad. Sci. U.S.A. 82, 7242-7246.

Menegatti, E., Guarneri, M., Bolognesi, M., Ascenzi, P. and Amiconi, G. (1987a). Role of the primary specificity subsite on the interaction between serine proteinases and low molecular weight substrates and inhibitors. In: Macromolecular Bio- recognition: Principles and Methods, (I. Chaiken, E. Chiancone, A. Fontana, and P. Neri, eds), pp. 101-115. The Humana Press, Clifton.

Menegatti, E., Guarneri, M., Bolognesi, M., Ascenzi, P. and Amiconi, G. (1987b). Binding of the bovine pancreatic secretory trypsin inhibitor (Kazal) to bovine serine (pro)enzymes. J. Mol. Biol. 198, 129-132.

Meyer, B. J., Badimon, J. J., Mailhac, A., FernBndez-Ortiz, A., Chesebro, J. H., Fuster, V. and Badimon, L. (1994a). Inhibition of growth of thrombus on fresh mural thrombus: targeting optimal therapy. Circulation 90, 2432-2438.

Meyer, B. J., FernBndez-Ortiz, A., Mailhac, A., Falk, E., Badimon, L., Don Michael, A., Chesebro, J. H., Fuster, V. and Badiman, J. J. (1994b). Local delivery of r-hirudin by a double-balloon perfusion catheter prevents mural thrombosis and minimizes platelet deposition after angioplasty. Circulation 90, 2474-2480.

Mignatti, P., Tsuboi, R., Robbins, E. and Rifkin, D. B. (1989). In vitro angiogenesis on the human amniotic membrane: requirement for basic fibroblast growth factor- induced proteinases. J. Cell. Biol. 108, 671-682.

Morrison, H. M., Welgus, H. G., Stockley, R. A., Burnett, D. and Campbell, E. J. (1990). Inhibition of human leukocyte elastase bound to elastin: relative ineffectiveness and two mechanisms of inhibitory activity. Am. J. Respir. Cell. Mol. Biol. 2, 263-269.

Miihlhahn, P., Czisch, M., Morenweiser, R., Habermann, B., Engh, R. A., Sommerhoff, C. P., Auerswald, E. A. and Holak, T. A. (1994). Structure of leech derived tryptase inhibitor (LDTI-c) in solution. FEBS Lett. 355, 29&296.

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.

Naski, M. C., Fenton, J. W. II, Maraganore, J. M., Olson, S. T. and Shafer, J. A. (1990). The COOH-terminal domain of hirudin: an exosite-directed competitive inhibitor of the action of cy-thrombin on fibrinogen. J. Biol. Chem. 265, 13484-13489.


Ni, F., Konishi, Y. and Scheraga, H. (1990). Thrombin-bound conformation of the C-terminal fragments of hirudin determined by transferred Nuclear Overhauser Effects. Biochemistry 29, 4479-4489.

Nicander of Colophain (1499). ,Theriaca et alexipharmaca. Apud Aldum Manutium, Venetiis (Modern edition and translation by Gowand, A. S. F. and Schofield, A. F. Cambridge, 1953).

Nick, H. P., Probst, A. and Schnebli, H. P. (1988). Development of eglin c as a drug: pharmacokinetics. Adv. Exp. Med. Biol. 240, 83-88.

Northrup, S. H. and Erickson, H. P. (1992). Kinetics of protein-protein association explained by Brownian dynamics computer simulation. Proc. Natl. Acad. Sci. U.S.A. 89, 3338-3342.

Okada, Y., Tsuboi, S., Nagamatsu, Y. and Yamamoto, J. (1990a). Synthesis of a heptacontapeptide corresponding to the entire amino acid sequence of eglin c. Chem. Pharm. Bull. Tokyo 38, 2902-2904.

Okada, Y., Tsuboi, S., Tsuda, Y., Nagamatsu, Y. and Yamamoto, J. (1990b). Synthesis of a trihexacontapeptide corresponding to the sequence 8-70 of eglin c and studies on the relationship between the structure and the inhibitory activity against human leukocyte elastase, cathepsin G and a-chymotrypsin. FEBS Lett. 272, 113-116.

Okada, Y., Tsuboi, S., Tsuda, Y., Nakabayashi, K., Nagamatsu, Y. and Yamamoto, J. (1989). Synthesis of peptide fragments related to eglin c and examination of their inhibitory effect on human leukocyte elastase, cathepsin G and a-chymotrypsin. Biochem. Biophys. Res. Commun. 161, 272-275.

Olson, S. T. and Shore, J. D. (1982). Demonstration of a two-step reaction mechanism for inhibition of a-thrombin by antithrombin III and identification of the step affected by heparin. J. Biol. Chem. 257, 14891-14895.

Oltersdorf, T., Fritz, L. C. and Schenck, D. (1989). The secreted form of the Alzheimer’s amyloid precursor protein with the Kunitz domain is protease nexin- II. Nature 341, 144-147.

Onesti, S., Matthews, D. J., Aducci, P., Amiconi, G., Bolognesi, M., Menegatti, E. and Ascenzi, P. (1992). Binding of the Kunitz-type trypsin inhibitor DE-3 from Erythrina caffra seeds to serine proteinases: a comparative study. J. Mol. Recogn. 5, 105-114.

Parry, M. A. A., Maraganore, J. M. and Stone, S. R. (1994). Kinetic mechanism for the interaction of hirulog with thrombin. Biochemistry 33, 14807-14814.

Patrignani, P., Del Maschio, A., Bazzoni, G., Daffonchio, L., Hernandez, A., Modica, R., Montesanti, L., Volpi, D., Patrono, C. and Dejana, E. (1991). Inactivation of endothelin by polymorphonuclear leukocyte-derived lytic enzymes. Bfood 78, 2715-2720.


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, 8571-8576.

Pintucci, G., Iacoviello, L., Castelli, M. P., Amore, C., Evangelista, V., Cerletti, C. and Donati, M. B. (1993). Capthepsin G-induced release of PAI-1 in the culture medium of endothelial cells: a new thrombotic role for polymorphonuclear leukocytes? J. Lab. Clin. Med. 122, 69-79.

Pizzo, S. V., Masts, A. E., Feldman, S. R. and Salvesen, G. (1988). In viva catabolism of cr,-antichymotrypsin is mediated by the serpin receptor which binds a,-proteinase inhibitor, antithrombin iii and heparin cofactor II. Biochim. Biophys. Acta 967, 158-162.

Prescott, S. M., Seeger, A. R., Zimmerman, G. A., Melutyre, T. M. and Maraganore, J. M. (1990). Hirudin-based peptides block the inflammatory effects of thrombin on endothelial cells. J. Biof. Chem. 265, 9614-9616.

Priestle, J. P., Rahuel, J., Rink, H., Tones, M. and Grutter, M. G. (1993). Changes in interactions in complexes of hirudin derivatives and human a-thrombin due to different crystal forms. Protein Sci. 2, 1630-1642.

Qiu, X., Padmanabhan, K. P., Carperos, V. E., Tulinsky, A., Kline, T., Maraganore, J. M. and Fenton, J. W. II (1992). Structure of the hirulog 3-thrombin complex and nature of the S’ subsites of substrates and inhibitors. Biochemistry 31, 11689-11697.

Qiu, X., Yin, M., Padmanabhan, K. P., Krstenansky, J. L. and Tulinsky, A. (1993). Structures of thrombin complexes with a designed and a natural exosite peptide inhibitor. J. Biol. Chem. 268, 20318-20326

Quast, U., Engel, J., Heumann, H., Krause, G. and Steffen, E. (1974). Kinetics of the interaction of bovine pancreatic trypsin inhibitor (Kunitz) with a-chymotrypsin. Biochemistry 13, 2512-2520.

Quast, U., Engel, J., Steffen, E., Tschesche, H. and Kupfer, S. (1978). Stopped-flow kinetics of the resynthesis of the reactive site peptide bond in kallikrein inhibitor (Kunitz) by l3-trypsin. Biochemistry 17, 1675-1682.

Rao, N. V., Wehner, N. G., Marshall, B. C., Gray, W. R., Gray, B. H., and Hoidal, J. R. (1991). Characterization of proteinase-3 (PR-3)) a neutrophil serine proteinase: structural and functional properties. J. Biol. Chem. 266, 9540-9546.

Read, R. J. and James. M. N. G. (1986). Introduction to the protein inhibitors: X-ray crystallography. In: Proteinme Inhibitors. (A. J. Barrett, and G. Salvesen, eds), pp. 301-336. Elsevier, Amsterdam, New York, Oxford.


Renesto, P. and Chignard, M. (1991). Tumour necrosis factor-a enhances platelet activation via cathepsin G released from neutrophils. J. Zmmunol. 146,2305-2309.

Renesto, P., Ferrer-Lopez, P. and Chignard, M. (1990). Interference of recombinant eglin c, a proteinase inhibitor extracted from leeches, with neutrophyl-mediated platelet activation. Lab. Invest. 62, 409-416.

Renesto, P., Ferrer-Lopez, P. and Chignard, M. (1991). Eglin c and heparin inhibition of platelet activation induced by cathepsin G or human neutrophils. Ann. N. Y. Acad. Sci. 624, 321-324.

Renesto, P., Halbwachs-Mecarelli, L., Nusbaum, P., Lesavre, P. and Chignard, M. (1994). Proteinase 3: a neutrophyl proteinase with activity on platelets. J. Zmmunol. 152, 4612-4617.

Repine, J. E. (1992). Scientific perspective on adult respiratory distress syndrome. Lancet 339, 466-469.

Ribeiro, J. M. C. and Garcia, E. S. (1981). Platelet antiaggregating activity in the salivary secretion of the blood-sucking bug Rhodnius prolixus. Experentia 37, 384-386.

Ribeiro, J. M. C., Makoul, G. T., Levine, J., Robinson, D. R. and Spielman, A. (1985). Antihemostatic, antiinflammatory and immunosuppressive properties of the saliva of the tick Zxodes dummini. J. Exp. Med. 161, 332-344.

Richardson, M. and Cossins, L. (1974). Chymotryptic inhibitor I from potatoes: the amino acid sequences of subunits B, C and D. FEBS Lett. 45, 11-13, and Corrigendum FEBS Lett. 52, 161.

Rigbi, M., Levy, H., Iraqi, F., Teitelbaum, M., Orevi, M., Alajoutsijarvi, A., Horovitz, A. and Galun R. (1987a). The saliva of the medicinal leech Hirudo medicinalis: 1. biochemical characterization of the high molecular weight fraction. Camp. Biochem. Physiol. 87, 567-573.

Rigbi, M., Levy, H., Elder, A., Iraqui, F., Teitelbaum, M., Orevi, M., Horovitz, A. and Galun, R. (1987b). The Saliva of the Medicinal Leech Hirudo medicinalis: II. Inhibition of platelet aggregation and of leukocyte activity and examination of reputed anaesthetic effects. Camp. Biochem. Physiol. C88, 95-98.

Ringe, D. and Petsko, G. A. (1985). Mapping protein dynamics by X-ray diffraction. Prog. Biophys. Mol. Biol. 45, 197-235.

Rink, K., Lierch, M., Sieber, P. and Meier, F. (1984). A large fragment approach to DNA synthesis: total synthesis of a gene for the protease inhibitor eglin c from Hirudo medicinalis and its expression in E.coli. Nucleic Acids Res. 13, 6369-6387.


Rosengren, S. and Arfors, K. E. (1990). Neutrophyl-mediated vascular leakage is not suppressed by leukocyte elastase inhibitors. Am. J. Physiol. 259, H128&H1294.

Roters, F. J. and Zebe, E. (1992). Protease inhibitors in the alimentary tract of the medicinal leech Hirudo medicinalis: in vivo and in vitro studies. J. Comp. Physiol. B162, 85-92.

Rydel, T. J., Ravichandran, K. G., Tulinsky, A., Bode, W., Huber, R., Roitsch, C. and Fenton, J. W. II (1990). The structure of a complex of recombinant hirudin and human o-thrombin. Science 249, 277-280.

Rydel, T. J., Tulinsky, A., Bode, W. and Huber, R. (1991). Refined structure of the hirudin-thrombin complex. J. Mol. Biol. 221, 583-601.

Scacheri, E., Nitti, G., Valsasina, B., Orsini, G., Visco, C., Ferrera, M., Sawyer, R. T. and Sarmientos, P. (1993). Novel hirudin variants from the leech Hirudinaria manillensis: amino acid sequence, cDNA cloning and genomic organization. Eur. J. Biochem. 214, 295-304.

Schar, H., Marki, W., Ghisalba, O., Jenny, H. and Rink, H. (1988). Modification of recombinant N-a-acetyl-eglin c by novel proteases from Nocardia mediterranei. Ann. N. Y. Acad. Sci. 542, 302-306.

Scharf, M., Engels, M. and Tripier, D. (1989). Primary structures of new ‘Iso-hirudins’. FEBS Lett. 255, 105-110.

Schechter, I. and Berger, A. (1967). On the size of the active site in proteases: I. papain. Biochem. Biophys. Res. Commun. 27, 157-162.

Schleuning, M., Schmidt-Haslbeck, M., Wimmer, R. and Mempel, W. (1990). Complement activation during storage of preserved blood. Beitr. Znfusionther. 26, 81-85.

Schleuning, M., Schmidt-Haslbeck, M., Utz, H., Jochum, M., Heim, M., Mempel, W. and Wilmanns, W. (1992). Complement activation during storage of blood under normal blood bank conditions: effects of proteinase inhibitors and leukocyte depletion. Blood 79, 3071-3075.

Schmitt, M., Kanayama, N., Henschen, A., Hollrieder, A., Hafter, R., Gulba, D., Janicke, F. and Graeff, H. (1989). Elastase released from human granulocytes stimulated with N-formyl-chemotactic peptide prevents activation of tumour cell prourokinase (pro-uPA). FEBS Lett. 255, 83-88.

Schmitt, M., Kanayama, N., Janicke, F., Hafter, R. and Graeff, H. (1991). Human tumour cell urokinase-type plasminogen activator (uPA): degradation of the proenzyme form (pro-uPA) by granulocyte elastase prevents subsequent activation by plasmin. Adv. Exp. Med. Biol. 297, 111-128.


Schmitz, T., Rothe, M. and Dodt, J. (1991). Mechanism of the inhibition of ol-thrombin by hirudin-derived fragments hirudin (l-47) and hirudin (45-65). Eur. J. Biochem. 195, 251-256.

Schnebli, H. P. and Braun, N. J. (1986). Proteinase inhibitors as drugs. In: Proteinuse Znhibitors (A. J. Barrett, and G. Salvesen, eds), pp. 613-627. Elsevier, Amsterdam, New York, Oxford.

Schreuder, H., Arkema, A., De Boer, B., Kalk, K., Dijkema, R., Mulders, J., Theunissen, H. J. M. and Hol, W. G. J. (1993). Crystallization and preliminary crystallographic analysis of antistasin, a leech-derived inhibitor of blood coagulation factor Xa. J. Mol. Biol. 231, 1137-1138.

Schreuder, H. A., De Boer, B., Dijkema, R., Mulders, J., Theunissen, H. J. M., Grootenhuis, P. D. J. and Hol, W. G. J. (1994). Crystal structure of a complex between intact and cleaved human antithrombin III as a model for serpin-proteinase interaction. Nature Struct. Biol. 1, 48-55.

Seemtiller, U., Dodt, J., Fink, E. and Fritz, H. (1986). Proteinase inhibitors of the leech Hirudo medicinalis (hirudins, bdellins, eglins). In: Proteinuse Inhibitors (A. J. Barrett and G. Salvesen, eds), pp. 337-359. Elsevier, Amsterdam, New York, Oxford.

Seemtiller, U., Meier, M., Ohlsson, K., Mtiller, H. P. and Fritz, H. (1977). Isolation and characterization of a low molecular weight inhibitor (of chymotrypsin and human granulocyte elastase and cathepsin G) from leechs. Hoppe-Seyler’s Z. Physiol. Chem. 358, 1105-1117.

Senior, R. M., Connolly, N. L., Cury, J. D., Welgus, H. G. and Campbell, R. J. (1989). Elastin degradation by human alveolar macrophages: a prominent role of metalloproteinase activity. Am. Rev. Respir. Dis. 139, 1251-1256.

Siebeck, M., Hoffmann, H., Weipert, J. and Spannagl, M. (1989). Therapeutic effects of the combination of two proteinase inhibitors in endotoxin shock of the pig. Prog. Clin. Biol. Res. 308, 937-943.

Siebeck, M., Hoffmann, H., Weipert, J. and Fritz, H. (1992a). Effect of the elastase inhibitor eglin c in porcine endotoxin shock. Circ. Shock 36, 174-179.

Siebeck, M., Spannagl, M., Hoffmann, H. and Fink, E. (1992b). Effects of protease inhibitors in experimental septic shock. Agents Actions Suppl. 38, 421-427.

Skrzypczak-Jankun, E., Carperos, V. E., Ravichandran, K. G., Tulinsky, A., Westbrook, M. and Maraganore, J. M. (1991). Structure of the hirugen and hirulog 1 complexes of cr-thrombin. J. Mol. Biol. 221, 1379-1393.

Smith, C. E., Stack, M. S. and Johnson, D. A. (1987). Ozone effects on inhibitors of human neutrophil proteinases. Arch. Biochem. Biophys. 253, 146155.


Snider, G. L., Stone, J. P., Lucey, E. C., Breuer, R., Calore, J. D., Seshadri, T., Catanese, A., Mascher, R. and Schnebli, H. P. (1985). Eglin c, a polypeptide derived from the medicinal leech, prevents human neutrophyl elastase-induced emphysema and bronchial secretory cell metaplasia in the hamster. Am. Rev. Respir. Dis. 132, 1155-1161.

Sollner, C., Mentele, R., Eckerskorn, C., Fritz, H. and Sommerhoff, C. P. (1994). Isolation and characterization of hirustasin, an antistasin-type serine-proteinase inhibitor from the medical leech Hirudo medicinalis. Eur. J. Biochem. 219, 937-943.

Sommerhoff, C. P., Siillner, C., Mentele, R., Piechottka, G. P., Auerswold, E. A. and Fritz, H. (1994). A Kazal-type inhibitor of human mast cell tryptase: isolation from the medical leech Hirudo medicinalis, characterization, and sequence analysis. Biol. Chem. Hoppe-Seyler 375, 685-694.

Sponer, M., Nick, H. P. and Schnebli, H. P. (1991). Different susceptibility of elastase inhibitors to inactivation by proteinases from Staphylococcus aureus and Pseudomonas aeruginosa. Biof. Chem. Hoppe-Seyler 372, 963-970.

Stack, M. S., Smith, C. E., Dean, W. L. and Johnson, D. A. (1988). Bronchial leukocyte elastase inhibitor: hydrodynamic properties and interaction with CX?- macroglobulin-bound elastase. Arch. Biochem. Biophys. 260, 400-407.

Starke, K. (1989). The beginnings of hirudin. TIPS 10, 99.

Steinmeyer, J. and Kalbhen, D. A. (1991). Influence of some natural and semisynthetic agents on elastase and cathepsin G from polymorphonuclear granulocytes. Arzneimittelforschung 41, 77-80.

Stone, S. R. (1995). Thrombin inhibitors: a new generation of antithrombotics. Trends Cardiovasc. Med., in press.

Stone, S. R. and Hofsteenge, J. (1986). Kinetics of the inhibition of thrombin by hirudin . Biochemistry, 25, 4622-4628.

Stone, S. R. and Maraganore, J. M. (1993). Hirudin and hirudin-based peptides. Meth. Enzymol. 223, 312-336.

Stone, S. R., Braun, P. J. and Hofsteenge, J. (1987). Identification of regions of a-thrombin involved in its interaction with hirudin. Biochemistry 26, 4617-4624.

Stone, S. R., Dennis, S. and Hofsteenge, J. (1989). Quantitative evaluation of the contribution of ionic interactions to the formation of the thrombin-hirudin complex. Biochemistry 28, 68576863.

Stubbs, M. T. and Bode, W. (1993a). A player of many parts: the spotlight falls on thrombin’s structure. Thromb. Res. 63, l-58.


Stubbs, M. T. and Bode, W. (1993b). Crystal structures of thrombin and thrombin complexes as a framework for antithrombotic drug design. Perspect. Drug Disc. Design 1, 431-452.

Stubbs, M. T. and Bode, W. (1994). Coagulation factors and their inhibitors. Curr. Opin. Struct. Biol. 4, 823-832.

Stubbs, M. T. and Bode, W. (1995a). The structural basis of thrombin inhibition by hirudin and serpins. In: Thrombogenesis and its Znhibition (Ofson and Buchanan, eds), CRC Press Inc., Boca Raton, in press.

Stubbs, M. T. and Bode, W. (1995b). The clot thickens: clues provided by thrombin structure. TIBS 20, 23-28.

Suter, S. and Chevallier, I. (1988). The effect of eglin c on the function of human neutrophils in vitro. Biol. Chem. Hoppe-Seyler 369, 573-578.

Suter, S., Schaad, U. B., Tegner, H., Ohlsson, K., Desgrandchamps, D. and Waldvogel. F. A. (1986). Levels of free granulocyte elastase in bronchial secretions from patients with cystic fibrosis: effect of antimicrobial treatment against Pseudomonas aeruginosa. J. Infect. Dis. 153, 902-909.

Suttorp, N., Nolte, A., Wilke, A. and Drenckhahn, D. (1993). Human neutrophil elastase increases permeability of cultured pulmonary endothelial cell monolayers. Znt. J. Microcirc. Clin. Exp. 13, 187-203.

Szyperski, T., Antuch, W., Schick, M., Betz, A., Stone, S. R. and Wiithrich, K. (1994). Transient hydrogen bonds identified on the surface of the NMR solution structure of hirudin. Biochemistry 33, 9303-9310.

Szyperski, T., Guntert, P., Stone, S. R. and Wiithrich, K. (1992a). Nuclear magnetic resonance solution structure of hirudin( l-5 1) and comparison with corresponding three- dimensional structures determined using the complete 65-residue hirudin polypeptide chain. J. Mol. Biol. 228, 1193-1205.

Szyperski, T., Gunteret, P., Stone, S. R., Tulinsky, A., Bode, W., Huber, R. and Wiithrich, K. (1992b). Impact of protein-protein contacts on the conformation of thrombin-bound hirudin studied by comparison with the nuclear magnetic resonance solution structure of hirudin(l-51). J. Mol. Biol. 228, 1206-1211.

Tapparelli, C., Metternich, R., Ehrhardt, C. and Cook, N. S. (1993). Synthetic low-molecular weight thrombin inhibitors: molecular design and pharmacological profile. TIPS 14, 366-376.

Teschner, M., Schaefer, R. M., Rudolf, C., Kulzer, P., Peter, G., Heidland, A. (1988). Eglin c fails to reduce catabolism in acutely uremic rats. Adv. Exp. Med. Biol. 240, 331-337.


Tian, W.-X. and Tsou, C.-L. (1982). Determination of the rate constant of enzyme modification by measuring the substrate reaction in the presence of the modifier. Biochemistry 21, 1028-1032.

Totani, L., Piccoli, A., Pellegrini, G., Di Santo, A. and Lorenzet, R. (1994). Polymorphonuclear leukocytes enhance release of growth factors by cultured endothelial cells. Arterioscler. Thromb. 14, 125-132.

Travis, J. and Fritz, H. (1991). Potential problems in designing elastase inhibitors for therapy. Am. Rev. Respir. Dis. 143, 1412-1415.

Travis, J. and Salvesen, G. (1983). Human plasma proteinase inhibitors. Annu. Rev. Biochem. 52, 655-709.

Travis, J., Owen, M., George, P., Carrell, R., Rosenberg, S., Hallewell, R. A. and Barr, P. J. (1985). Isolation and properties of recombinant DNA produced variants of human a,-proteinase inhibitor. 1. Bioof. Chem. 260, 4384-4389.

Tripier, D. (1988). Hirudin: a family of iso-proteins: isolation and sequence determination of new hirudins. Folia Haematol. Leipz. 115, 3&35.

Tsuda, Y., Cygler, M., Gibbs, B. F., ‘Pedyczak, A., F&hi&e, J., Yue, S. Y. and Konishi, Y. (1994). Design of potent bivalent thrombin inhibitors based on hirudin sequence: incorporation of nonsubstrate-type active site inhibitors. Biochemistry 33, 14443-14451.

van Obberghen-Schilling, E., Perez-Rodriguez, R. and Pouyssegur, J. (1982). Hirudin, a probe to analyze the growth-promoting activity of thrombin in fibroblasts: reevaluation of the temporal action of competence factors. Biochem. Biophys. Res. Commun. 106, 79-86.

Verlinde, C. L. M. J. and Hol, W. G. J. (1994). Structure-based drug design: progress, results and challenges. Structure 2, 577-587.

Vijayalakshmi, J., Padmanabhan, K. P., Mann, K. G. and Tulinsky, A. (1994). The isomorphous structures of prethrombin 2, hirugen- and PPACK-Thrombin: changes accompanying activation and exosite binding to thrombin. Protein Sci. 3, 2254-2271.

Virca, G. D. and Travis, J. (1990). Kinetics of association of human proteinases with human u,-macroglobulin. J. Biol. Chem. 259, 8870-8874.

Virca, G. D., Metz, G. and Schnebli, H. P. (1984). Similarities between human and rat leukocyte elastase and cathepsin G. Eur. J. Biochem. 144, l-9.

Vitali, J., Martin, P. D., Malkowski, M. G., Robertson, W. D., Lazar, J. B., Winant, R. C., Johnson, P. J. and Edwards, B. F. P. (1992). The structure of a complex of bovine cy-thrombin and recombinant hirudin at 2.8 8, resolution. J. Biol. Chem. 267, 17670-17678.


von Ritter, C., Be, R. and Granger, D. N. (1989). Neutrophilic proteases: mediators of formyl-methionyl-leucyl-phenylalanine-induced ileitis in rats. Gustroenterology 97,605-609.

Wagner, G., Thanabal, W., Stockman, B. J., Peng, J. W., Nirmala, N. R., Hyberts, S. G., Goldberg, M. S., Detlefsen, D. J., Clubb, R. T. and Adler, M. (1992). NMR studies of structure and dynamics of isotope enriched proteins. Biopolymers 32, 381-390.

Walenga, J. M., Pifarre, R. and Fareed, J. (1990). Recombinant hirudin as antithrombotic agent. Drugs Fut. 15, 267-280.

Wallace, A., Dennis, S., Hofsteenge, J. and Stone, S. R. (1989). Contribution of the N-terminal region of hirudin to its interaction with thrombin. Biochemistry 28, 10079-10084.

Wallis, R. B. (1988). Hirudins and the role of thrombin: lessons from leechs. TIPS. 9, 425-427.

Wallis, R. B. (1989). Inhibition of thrombin, a key step in thrombosis. Drugs Today 25, 597-605.

Walsmann, P. and Markwardt , F. (1981). Biochemische und pharmakologische aspekte des trombininhibitors hirudin. Pharmuzie 36, 653-660.

Walsmann, P. and Kaiser, B. (1989). Biochemical and pharmacological properties of recombinant hirudin. Drugs Today 25, 473-485.

Weis, M. S. (1992). Der Einfluss von r-Hirudin, einem Spezifischen Thrombininhibitor, auf das Akute Lungenversagen im Septischen Schock am Tiermodell. Ph. D. Thesis, Univ. Mtinchen.

Weitz, J. I., Hudoba, M., Massel, D. R., Maraganore, J. and Hirsh, J. (1990). Clot-bound thrombin is protected from inhibition by heparin-antithrombin III but is susceptible to inactivation by antithrombin III-independent inhibitors. 1. Clin. Invest. 86, 385-391.

Wells, J. A. and Lowman, H. B. (1992). Rapid evolution of peptide and protein binding properties in vitro. Curr. Opin. Struct. Biol. 2, 597-604.

Wells, T. N. C. and Fersht, A. R. (1986). Use of binding energy in catalysis analyzed by mutagenesis of the tyrosyl-tRNA synthetase. Biochemistry 25, 1881-1886.

Wells, M. D., Manktelow, R. T., Boyd, J. B. and Bowen, V. (1993). The medical leech: an old treatment revisited. Microsurgery 14, 183-186.

Weng, M. (1993). Experimentaltherapeutischer Einsatzder Protea seninhibitoren Hirudin und Eglin sowie von Atheplex bei der Endotoxininduzierten Schocklunge des Hausschweines. Ph. D. Thesis, Univ. Freiburg (Breisgau).


Wewers, M. D., Herzyk, D. J. and Gadek, J. E. (1988). Alveolar fluid neutrophil elastase activity in the acute respiratory distress syndrome is complexed to cY,-macroglobulin. J. C/in. Invest. 82, 1260-1267.

Wiedow, O., Ludemann, J. and Utecht, B. (1991). Elafin is a potent inhibitor of proteinase 3. Biochem. Biophys. Res. Commun. 174, 6-10.

Wiedow, O., Liidemann, J., Utecht, B. and Christophers, E. (1993). Inhibition of proteinase 3 activity by peptides derived from human epidermis. Adv. Exp. Med. Biol. 336, 61-64.

Wilcox, J. N., Smith, K. M., Schwartz, S. M. and Gordon, D. (1989). Localization of tissue factor in the normal vessel wall and in the atherosclerotic plaque. Proc. Natl. Acad. Sci. U.S.A. 86, 2839-2843.

Winant, R. C., Lazar, J. B. and Johnson, P. H. (1991). Chemical modifications and amino acid substitutions in recombinant hirudin that increase hirudin-thrombin affinity. Biochemistry 30, 1271-1277.

Witting, J. I., Bourdon, P., Brezniak, D. V., Maraganore, J. M. and Fenton, J. W. II (1992). Thrombin-specific inhibition by and slow cleavage of hirulog-1. Biochem. J. 283, 737-743.

Woodman, R. C., Reinhardt, P. H., Kanwar, S., Johnstone, F. L. and Kubes, P. (1993). Effect of human neutrophil elastase (HNE) on neutrophil function in vitro and in inflamed microvessels. Blood 82, 2188-2195.

Yamaguchi, T., Kido, H., Fukase, M., Fujita, T. and Katunuma, M. (1991). A membrane-bound metallo-endopeptidase from rat kidney hydrolyzing parathyroid hormone. Eur. J. Biochem. 200, 563-571.

Yamamoto, J., Ishii, I., Okada, Y., Yamashita, T., Ridler, C. D., Gorog, P. and Kovacs, I. B. (1993). Effect of leukocyte products on platelet thrombus formation, coagulation and spontaneous thrombolysis, as measured from native human blood, in vitro. Thromb. Res. 71, 281-287.

Zdanov, A., Wu, S., DiMaio, J., Konishi, Y., Li, Y., Wu, X., Edwards, B. F., Martin, P. D. and Cygler, M. (1993). Crystal structure of the complex of human cy-thrombin and nonhydrolyzable bifunctional inhibitors, hirutonin-2 and hirutonin-6. Proteins 17, 252-265.

Zimmerman, B. J. and Granger, D. N. (1990). Reperfusion-induced leukocyte infiltration: role of elastase. Am. J. Physiol. 259, H390-H394.

Proteinase inhibitors from the european medicinal leech Hirudo medicinalis: Structural, functional and biomedical aspects

Documents