Structure of the complex formed by bovine trypsin and bovine pancreatic trypsin inhibitor. II. Crystallographic refinement at 1.9 A resolution

Crystal structure of an ecotin-collagenase complex suggests a model for recognition and cleavage of the collagen triple helix

The crystal structure of fiddler crab collagenase complexed with the dimeric serine protease inhibitor ecotin at 2.5 A resolution reveals an extended cleft providing binding sites for at least 11 contiguous substrate residues. Comparison of the positions of nine intermolecular main chain hydrogen bonding interactions in the cleft, with the known sequences at the cleavage site of type I collagen, suggests that the protease binding loop of ecotin adopts a conformation mimicking that of the cleaved strand of collagen. A well-defined groove extending across the binding surface of the enzyme readily accommodates the two other polypeptide chains of the triple-helical substrate. These observations permit construction of a detailed molecular model for collagen recognition and cleavage by this invertebrate serine protease. Ecotin undergoes a pronounced internal structural rearrangement which permits binding in the observed conformation. The capacity for such rearrangement appears to be a key determinant of its ability to inhibit a wide range of serine proteases.

Electrostatic complementarity at protein/protein interfaces

Calculation of the electrostatic potential of protein-protein complexes has led to the general assertion that protein-protein interfaces display "charge complementarity" and "electrostatic complementarity". In this study, quantitative measures for these two terms are developed and used to investigate protein-protein interfaces in a rigorous manner. Charge complementarity (CC) was defined using the correlation of charges on nearest neighbour atoms at the interface. All 12 protein-protein interfaces studied had insignificantly small CC values. Therefore, the term charge complementarity is not appropriate for the description of protein-protein interfaces when used in the sense measured by CC. Electrostatic complementarity (EC) was defined using the correlation of surface electrostatic potential at protein-protein interfaces. All twelve protein-protein interfaces studied had significant EC values, and thus the assertion that protein-protein association involves surfaces with complementary electrostatic potential was substantially confirmed. The term electrostatic complementarity can therefore be used to describe protein-protein interfaces when used in the sense measured by EC. Taken together, the results for CC and EC demonstrate the relevance of the long-range effects of charges, as described by the electrostatic potential at the binding interface. The EC value did not partition the complexes by type such as antigen-antibody and proteinase-inhibitor, as measures of the geometrical complementarity at protein-protein interfaces have done. The EC value was also not directly related to the number of salt bridges in the interface, and neutralisation of these salt bridges showed that other charges also contributed significantly to electrostatic complementarity and electrostatic interactions between the proteins. Electrostatic complementarity as defined by EC was extended to investigate the electrostatic similarity at the surface of influenza virus neuraminidase where the epitopes of two monoclonal antibodies, NC10 and NC41, overlap. Although NC10 and NC41 both have quite high values of EC for their interaction with neuraminidase, the similarity in electrostatic potential generated by the two on the overlapping region of the epitopes is insignificant. Thus, it is possible for two antibodies to recognise the electrostatic surface of a protein in dissimilar ways.

Rational design of complex formation between plasminogen activator inhibitor-1 and its target proteinases

Considerable progress in understanding the mechanism of inhibition of proteinases by serpins has been obtained from different biochemical studies. These studies reveal that stable serpin/proteinase complex formation involves insertion of the reactive-site loop of the serpin and occurs at the acyl-enzyme stage. Even though no three-dimensional structure of a serpin/proteinase complex is resolved, structural information is available on some of the individual compounds. Molecular modeling techniques combined with recently acquired biochemical/biophysical data were used to provide insight into the stable complex formation between plasminogen activator inhibitor-1 (PAI-1) and the target proteinases: tissue-type plasminogen activator, urokinase-type plasminogen activator, and thrombin. This study reveals that PAI-1 initially interacts with its target proteinase when its reactive-site loop is solvent exposed and thereby accessible for the proteinase. Stable complex formation, however, involves the insertion of the reactive-site loop up to P7 and results in a tight binding geometry between PAI-1 and its target proteinase. The influence of different biologically relevant molecules on PAI-1/proteinase complex formation and the differences in inhibition rate constants observed for the different proteinases can be explained from these models.

Modeling supra-molecular helices: extension of the molecular surface recognition algorithm and application to the protein coat of the tobacco mosaic virus

Geometric matching of molecular surfaces appears to be essential for the formation of binary molecular complexes and of supra-molecular aggregates. The structure of a binary complex is characterized by the best geometric match, whereas the structure of an aggregate is characterized by the best combined match, i.e. the sum of all the internal matches in the system. We describe a method to identify and quantify the binary matches between molecules and then use them to form the supra-molecular helices and evaluate them. This method is applied to the single protein subunit of tobacco mosaic virus. It successfully predicts the structure of the helical protein coat of the virus and the structure of the disk that is formed as the initial step in the virus assembly process. It also predicts structural intermediates, between disk and helix, which explain how the disk can transform into a helix without dissociating into subunits.

Enzyme-inhibitor association thermodynamics: explicit and continuum solvent studies

Studying the thermodynamics of biochemical association reactions at the microscopic level requires efficient sampling of the configurations of the reactants and solvent as a function of the reaction pathways. In most cases, the associating ligand and receptor have complementary interlocking shapes. Upon association, loosely connected or disconnected solvent cavities at and around the binding site are formed. Disconnected solvent regions lead to severe statistical sampling problems when simulations are performed with explicit solvent. It was recently proposed that, when such limitations are encountered, they might be overcome by the use of the grand canonical ensemble. Here we investigate one such case and report the association free energy profile (potential of mean force) between trypsin and benzamidine along a chosen reaction coordinate as calculated using the grand canonical Monte Carlo method. The free energy profile is also calculated for a continuum solvent model using the Poisson equation, and the results are compared to the explicit water simulations. The comparison shows that the continuum solvent approach is surprisingly successful in reproducing the explicit solvent simulation results. The Monte Carlo results are analyzed in detail with respect to solvation structure. In the binding site channel there are waters bridging the carbonyl oxygen groups of Asp189 with the NH2 groups of benzamidine, which are displaced upon inhibitor binding. A similar solvent-bridging configuration has been seen in the crystal structure of trypsin complexed with bovine pancreatic trypsin inhibitor. The predicted locations of other internal waters are in very good agreement with the positions found in the crystal structures, which supports the accuracy of the simulations.

Probing intermolecular main chain hydrogen bonding in serine proteinase-protein inhibitor complexes: chemical synthesis of backbone-engineered turkey ovomucoid third domain

Intermolecular main chain H-bonding networks are frequently encountered at the interface of complexes of protein proteinase inhibitors and their cognate enzymes. Studies of X-ray crystal structures of many protein inhibitors complexed with serine proteinases have revealed that the amide NH group of the P1 residue in the inhibitor donates an H-bond to the carbonyl C = O group of Ser214 and Ser195 Oy in the enzyme (Ser125 and Ser221 in subtilisins, respectively). To probe the energetic contribution of this backbone H-bond in the complexes of OMTKY3 with several serine proteinases, native chemical ligation was used for the total synthesis of a backbone-engineered analog of OMTKY3, in which the amide peptide bond between Thr17 (P2) and Leu18 (P1) was replaced by an ester bond, i.e., -CONH-to-COO-. This chemical "mutation" effectively eliminated the backbone H-bond donated by the NH group of Leu18. By measuring association equilibrium constants for synthetic wild-type OMTKY3 and the backbone-engineered ester analog interacting with a panel of six serine proteinases, we have determined that the P1 NH-->O substitution weakens the binding of OMTKY3 to its cognate enzymes by an average of 15-fold, i.e., 1.5 +/- 0.3 kcal/mol. These results place a quantitative value on the contribution of the intermolecular backbone H-bond in enzyme-inhibitor recognition.

All in one: a highly detailed rotamer library improves both accuracy and speed in the modelling of sidechains by dead-end elimination

BACKGROUND

About a decade ago, the concept of rotamer libraries was introduced to model sidechains given known mainchain coordinates. Since then, several groups have developed methods to handle the challenging combinatorial problem that is faced when searching rotamer libraries. To avoid a combinatorial explosion, the dead-end elimination method detects and eliminates rotamers that cannot be members of the global minimum energy conformation (GMEC). Several groups have applied and further developed this method in the fields of homology modelling and protein design.

RESULTS

This work addresses at the same time increased prediction accuracy and calculation speed improvements. The proposed enhancements allow the elimination of more than one-third of the possible rotameric states before applying the dead-end elimination method. This is achieved by using a highly detailed rotamer library allowing the safe application of an energy-based rejection criterion without risking the elimination of a GMEC rotamer. As a result, we gain both in modelling accuracy and in computational speed. Being completely automated, the current implementation of the dead-end elimination prediction of protein sidechains can be applied to the modelling of sidechains of proteins of any size on the high-end computer systems currently used in molecular modelling. The improved accuracy is highlighted in a comparative study on a collection of proteins of varying size for which score results have previously been published by multiple groups. Furthermore, we propose a new validation method for the scoring of the modelled structure versus the experimental data based upon the volume overlap of the predicted and observed sidechains. This overlap criterion is discussed in relation to the classic RMSD and the frequently used +/- 40 degrees window in comparing chi 1 and chi 2 angles.

CONCLUSIONS

We have shown that a very detailed library allows the introduction of a safe energy threshold rejection criterion, thereby increasing both the execution speed and the accuracy of the modelling program. We speculate that the current method will allow the sidechain prediction of medium-sized proteins and complex protein interfaces involving up to 150 residues on low-end desktop computers.

Crystal structure of the bovine alpha-chymotrypsin:Kunitz inhibitor complex. An example of multiple protein:protein recognition sites

The crystal structure of bovine alpha-chymotrypsin (alpha-CHT) in complex with the bovine basic pancreatic trypsin inhibitor (BPTI) has been solved and refined at 2.8 A resolution (R-factor = 0.18). The proteinase:inhibitor complex forms a compact dimer (two alpha-CHT and two BPTI molecules), which may be stabilized by surface-bound sulphate ions, in the crystalline state. Each BPTI molecule, at opposite ends, is contacting both proteinase molecules in the dimer, through the reactive site loop and through residues next to the inhibitor's C-terminal region. Specific recognition between alpha-CHT and BPTI occurs at the (re)active site interface according to structural rules inferred from the analysis of homologous serine proteinase:inhibitor complexes. Lys15, the P1 residue of BPTI, however, does not occupy the alpha-CHT S1 specificity pocket, being hydrogen bonded to backbone atoms of the enzyme surface residues Gly216 and Ser217.

Hypothesis for a serine proteinase-like domain at the COOH terminus of Slowpoke calcium-activated potassium channels

Bovine pancreatic trypsin inhibitor (BPTI) is a 58-residue protein with three disulfide bonds that belongs to the Kunitz family of serine proteinase inhibitors. BPTI is an extremely potent inhibitor of trypsin, but it also specifically binds to various active and inactive serine proteinase homologs with KD values that range over eight orders of magnitude. We previously described an interaction of BPTI at an intracellular site that results in the production of discrete subconductance events in large conductance Ca2+ activated K+ channels (Moss, G.W.J., and E. Moczydlowski. 1996, J. Gen. Physiol, 107:47-68). In this paper, we summarize a variety of accumulated evidence which suggests that BPTI binds to a site on the KCa channel protein that structurally resembles a serine proteinase. One line of evidence includes the finding that the complex of BPTI and trypsin, in which the inhibitory loop of BPTI is masked by interaction with trypsin, is completely ineffective in the production of substate events in the KCa channel. To further investigate this notion, we performed a sequence analysis of the alpha-subunit of cloned slowpoke KCa channels from Drosophila and mammals. This analysis suggests that a region of approximately 250 residues near the COOH terminus of the KCa channel is homologous to members of the serine proteinase family, but is catalytically inactive because of various substitutions of key catalytic residues. The sequence analysis also predicts the location of a Ca(2+)-binding loop that is found in many serine proteinase enzymes. We hypothesize that this COOH-terminal domain of the slowpoke KCa channel adopts the characteristic double-barrel fold of serine proteinases, is involved in Ca(2+)-activation of the channel, and may also bind other intracellular components that regulate KCa channel activity.

Strategy to design peptide inhibitors: structure of a complex of proteinase K with a designed octapeptide inhibitor N-Ac-Pro-Ala-Pro-Phe-DAla-Ala-Ala-Ala-NH2 at 2.5 A resolution

The crystal structure of a complex formed by the interaction between proteinase K and a designed octapeptide amide, N-Ac-Pro-Ala-Pro-Phe-DAla-Ala-Ala-Ala-NH2, has been determined at 2.5 A resolution and refined to an R-factor of 16.7% for 7,430 reflections in the resolution range of 8.0-2.50 A. The inhibitor forms a stable complex through a series of hydrogen bonds and hydrophobic interactions with the protein atoms and water molecules. The inhibitor is hydrolyzed between Phe4I and DAla5I (I indicates the inhibitor). The two fragments are separated by a distance of 3.2 A between the carbonyl carbon of Phe4I and the main-chain nitrogen of DAla5I. The N-terminal tetrapeptide occupies subsites S1-S5 (S5 for acetyl group), whereas the C-terminal part fits into S1'-S5' region (S5' for amide group). It is the first time that such an extended electron density for a designed synthetic peptide inhibitor has been observed in the prime region of an enzyme of the subtilisin family. In fact, the inhibitor fills the recognition site completely. There is only a slight rearrangement of the protein residues to accommodate the inhibitor. Superposition of the present octapeptide inhibitor on the hexapeptide inhibitor studied previously shows an overall homology of the two inhibitors, although the individual atoms are displaced significantly. It suggests the existence of a recognition site with flexible dimensions. Kinetic studies indicate an inhibition rate of 100% by this specifically designed peptide inhibitor.

X-ray structure of active site-inhibited clotting factor Xa. Implications for drug design and substrate recognition

The 3.0-A resolution x-ray structure of human des-Gla-coagulation factor Xa (fXa) has been determined in complex with the synthetic inhibitor DX-9065a. The binding geometry is characterized primarily by two interaction sites: the naphthamidine group is fixed in the S1 pocket by a typical salt bridge to Asp-189, while the pyrrolidine ring binds in the unique aryl-binding site (S4) of fXa. Unlike the large majority of inhibitor complexes with serine proteinases, Gly-216 (S3) does not contribute to hydrogen bond formation. In contrast to typical thrombin binding modes, the S2 site of fXa cannot be used by DX-9065a since it is blocked by Tyr-99, and the aryl-binding site (S4) of fXa is lined by carbonyl oxygen atoms that can accommodate positive charges. This has implications for natural substrate recognition as well as for drug design.

An investigation of the binding of protein proteinase inhibitors to trypsin by electrospray ionization mass spectrometry

The binding of BPTI and SBTI with trypsin has been investigated by ESI MS, using the mutant K15V-BPTI and the chemically modified RcamBPTI as controls. Although high cone voltages (+80 V) produce sharp spectra of BPTI, RcamBPTI, SBTI and trypsin alone, the complexes of BPTI, RcamBPTI and SBTI with trypsin undergo partial dissociation due to collisional activation. At lower cone voltages (+40 V) these non-covalent complexes are stable. The charge distribution on the trypsin and the inhibitors produced by gas phase dissociation of the complexes are markedly different from those of the components alone, indicating that ESI MS provides a novel probe for exploring the ionic interactions at the contact surface of proteins. Moreover, by determining the cone voltage at which the gas phase dissociation of complexes occurs it may be possible to use ESI MS to compare the binding energies of closely related complexes.

Three-dimensional structure of human cytomegalovirus protease

Herpesviruses encode a serine protease that specifically cleaves assembly protein. This protease is critical for replication, and represents a new target for antiviral drug design. Here we report the three-dimensional structure of the protease from human cytomegalovirus (hCMV) at 2.27 angstroms resolution. The structure reveals a unique fold and new catalytic strategy for cleavage. The monomer fold of the enzyme, a seven-stranded beta-barrel encircled by a chain of helices that form the carboxy terminus of the molecule, is unrelated to those observed in classic serine proteases such as chymotrypsin and subtilisin. The serine nucleophile at position 132 is activated by two juxtaposed histidine residues at positions 63 and 157. Dimerization, which seems to be necessary for activity, is observed in the crystals. Correlations of the structure with the sequences of herpesvirus proteases suggest that dimerization may confer specificity and recognition in substrate binding.

Alanine point-mutations in the reactive region of bovine pancreatic trypsin inhibitor: effects on the kinetics and thermodynamics of binding to beta-trypsin and alpha-chymotrypsin

In an effort to relate structural, kinetic, and thermodynamic features in a model macromolecular recognition process, the amino acid residues in the reactive surface of bovine pancreatic trypsin inhibitor (BPTI) and surrounding residues were substituted individually by alanine, and the effects of the point-mutations on the kinetics and thermodynamics of inhibition by BPTI toward trypsin and chymotrypsin were investigated. Fifteen alanine mutants were produced. The majority of the BPTI mutants exhibited a binding affinity similar to that of the wild-type protein. The exceptions were the primary specificity site (PI) mutant and those mutants that seem to have nonlocal perturbations of structure, as revealed by circular dichroism and thermostability measurements. The mutation at the P1 site caused a reduction in the binding free energy of 10 and 1.8 kcal mol-1 for trypsin and chymotrypsin, respectively. The losses in binding affinity were determined almost exclusively by an increase in the dissociation rate constant. However, the rate of association of the P1 mutant, Lys-15-Ala, with trypsin was also drastically reduced (> 200-fold). Calorimetric measurements of the heats of binding for the association of chymotrypsin with the wild-type inhibitor and its alanine mutants allowed determination of the relative contributions of the changes in enthalpy and entropy to the free energy of binding. Compensatory changes in the two parameters were observed in several cases, which were attributed to desolvation effects at the binding interface.

Role of the P2 residue in determining the specificity of serpins

The importance of the P2 residue in determining serpin specificity was examined by making a series of substitutions in the P2 position of recombinant alpha 1-antichymotrypsin that contained an arginine P1 residue. The importance of the P2 residue in governing the association rate constant (Kon) of the serpin varied with the protease examined. For trypsin, the P2 residue played a relatively minor role, whereas the nature of this residue markedly influenced the rates of inhibition of thrombin, factor Xa, and APC. A 1000-fold difference in Kon values was observed between the fastest (P2 proline) and the slowest (P2 threonine) inhibitors of thrombin. Similar differences were observed with factor Xa; the best inhibitor (P2 glycine) displayed a 200-fold higher Kon value than the poorest (P2 threonine). The nature of the P2 residue also affected whether the interaction of the serpin with the protease resulted in inhibition of the protease or cleavage of the serpin; a P2 proline residue increased the rate of cleavage of alpha 1-antichymotrypsin by trypsin. By using mutants of thrombin, it was possible to show that the B-insertion loop, which partially occludes the active site, is important in determining the P2 specificity of this enzyme. Deletion of three amino acids from this loop yielded a protease (des-PPW) that became more like trypsin in its specificity. In addition, it was shown that Glu192 dramatically restricts thrombin's ability to accommodate a threonine in the P2 position. Taken together, the results demonstrated the importance of complementary interactions between the P2 residue of the serpin and the S2 binding site of the protease in regulating the specific interaction between serpin and protease.

The design of an inhibitor using alpha, beta-dehydro residues

X-ray crystallography and NMR spectroscopic studies on peptides containing alpha, beta-unsaturated (dehydro) residues have indicated that the steric effects caused by the dehydro residue are strong and predictable, and such residues can be used to generate specific peptide structures. In particular, dehydro-Alanine (delta Ala) is found to adopt an extended conformation and also induces a definite conformation in the neighboring saturated residue. In order to design a sequence that would fold into a known three-dimensional structure, we have undertaken a systematic theoretical study of the preferred conformations of tetrapeptide sequences of the type N-Ac-delta Ala-X-delta Ala-NHCH3 (x = Gly, L-Ala, L-Val, L-Leu, L-Lys, L-Arg, L-Phe). The methodology and parameters used have been standardized against sequences with known crystal structures. In all the sequences a consistent folding pattern is observed in which the delta Ala residues are in an extended conformation with phi approximately 140 degrees and psi approximately -40 degrees. The results are used to predict a sequence which has a structure very similar to that of the binding loop region of the bovine pancreatic trypsin inhibitor (BPTI).

Inhibition of human leukocyte and porcine pancreatic elastase by homologues of bovine pancreatic trypsin inhibitor

The interactions of three BPTI homologues with human leukocyte elastase and porcine pancreatic elastase have been investigated. The principal mutation in determining the specificity of inhibition was the Lys15-Val mutation at the P1 position. An additional mutation at P3, i.e., BPTI (Lys15-Val, Pro13-Ile), increased the inhibition of HLE to a Ki = 2.5 x 10(-10) M, but decreased the inhibition of PPE, showing this to be a useful site for improving selectivity. Kinetic evidence suggests that the inhibition of HLE by BPTI homologues probably takes place by a two-step mechanism in which an isomerization step occurs after initial binding. 1H NMR spectroscopy of the BPTI (Lys15-Val) and BPTI (Lys15-Val, Pro13-Ile) mutants indicates that small conformational changes are associated with the mutations, but these are localized in the immediate vicinity of the mutation in the outer binding loop and in the inner loop connected to it through the Cys14-Cys38 disulfide bridge.

Stable monomeric form of an originally dimeric serine proteinase inhibitor, ecotin, was constructed via site directed mutagenesis

Ecotin, a homodimer protein of E. coli, is a unique member of canonical serine proteinase inhibitors, since it is a potent agent against a variety of serine proteinases having different substrate specificity. Monomers of ecotin are held together mostly by their long C-terminal strands that are arranged as a two-stranded antiparallel beta-sheet in the functional dimer. One ecotin dimer can chelate two proteinase molecules, each of them bound to both subunits of ecotin at two different sites, namely the specific primary and the non-specific secondary binding sites. In this study the genes of wild type ecotin and its Met84Arg P1 site mutant were truncated resulting in new forms of ecotin that lack 10 amino acid residues at their C-terminus. These mutants do not dimerize spontaneously, though in combination with trypsin they assemble into the familiar heterotetramer. Our data suggest that this heterotetramer exists even in extremely diluted solutions, and the interaction, which is responsible for the dimerization of ecotin, contributes to the stability of the heterotetrameric complex.

Inhibitory properties of separate recombinant Kunitz-type-protease-inhibitor domains from tissue-factor-pathway inhibitor

Tissue-factor-pathway inhibitor (TFPI) is a multivalent inhibitor with three tandemly arranged Kunitz- type-protease-inhibitor (KPI) domains. Previous studies [Girard, Y. J., Warren, L. A., Novotny , W. F., Likert, K. M., Brown, S. G., Miletich, J. R & Broze, G. J. (1989) Nature 338, 518-520] by means of site-directed mutagenesis indicated that KPI domain 1 interacts with factor VIIa, that KPI domain 2 interacts with factor Xa, and that KPI domain 3 is apparently without inhibitory function. To elucidate the reaction mechanism of this complex inhibitor, we followed a different approach and studied the inhibitory properties of fragments of TFPI obtained by expression in yeast. Results obtained with TFPI-(1-161)-peptide and separate recombinant TFPI-KPI domains 1, 2 and 3 showed that KPI domain 1 inhibited factor VIIa/tissue factor (Ki = 250 nM), KPI domain 2 inhibited factor Xa (Ki = 90 nM), and that KPI domain 3 was without detectable inhibitory function. Studies with separate KPI domains also showed that KPI domain 2 was mainly responsible for inhibition of trypsin (Ki = 0.1 nM) and chymotrypsin (Ki = 0.75 nM), whereas KPI domain 1 inhibited plasmin (Ki = 26 nM) and cathepsin G (Ki = 200 nM). The structural basis for the interaction between serine proteases and KPI domains is discussed in terms of putative three-dimensional models of the proteins derived by comparative molecular-modelling methods. Studies of factor Xa inhibition by intact TFPI (Ki approximately 0.02 nM) suggested that regions other than the contact area of the KPI domain, interacted strongly with factor Xa. Secondary-site interactions were crucial for TFPI inhibition of factor Xa but was of little or no importance for its inhibition of trypsin.

Synthesis of a mixture of cyclic peptides based on the Bowman-Birk reactive site loop to screen for serine protease inhibitors

A peptide mixture containing 21 peptide sequences has been constructed to test the Bowman-Birk inhibitor reactive-site loop motif as the basis of inhibition for a range of serine proteases. The 21 peptides are all based on an 11 amino acid sequence designed from a Bowman-Birk like inhibitor reactive-site loop. Variation has been introduced at the P1 site of the loop, which has been randomised to include all the natural L-amino acids (except for cysteine), plus the non-natural L-amino acids ornithine and norleucine, The mixture of peptides was screened for specific binding to immobilised porcine pancreatic elastase, subtilisin BPN', alpha-chymotrypsin, trypsin, anhydro-alpha-chymotrypsin and anhydrotrypsin. Five peptides from the mixture bind to alpha-chymotrypsin, two of which also bind to anhydro-alpha-chymotrypsin, and two peptides bind trypsin, neither of which binds to anhydro-trypsin. The competitive inhibition constants (K(i)) and the rates of proteolytic hydrolysis of the individual peptides with their respective enzymes were determined. The rates of hydrolysis were found to vary widely and show little correlation with the K(i) values. In the case of the alpha-chymotrypsin inhibitors, the peptides with the lowest K(i) (0.1-0.05 mM) were the only peptides that bound to anhydro-alpha-chymotrypsin. However, no peptides bound to anhydrotrypsin, suggesting a fundamental difference in the way that alpha-chymotrypsin and trypsin are inhibited by these cyclic peptides.

Elusive affinities

The affinity of two molecules for each other and its temperature dependence are determined by the change in enthalpy, free enthalpy, entropy, and heat capacity upon dissociation. As we know the forces that stabilize protein-protein or protein-DNA association and the three-dimensional structures of the complex, we can in principle derive values for each one of these parameters. The calculation is done first in gas phase by molecular mechanics, then in solution with the help of hydration parameters calibrated on small molecules. However, estimates of enthalpy and entropy changes in gas phase have excessively large error bars even under the approximation that the components of the complex associate as rigid bodies. No reliable result can be expected at the end. The fit to experimental values derived from binding and calorimetric measurements is poor, except for the dissociation heat capacity. This parameter can be attributed mostly to the hydration step and it correlates with the size of the interface. Many protein-protein complexes have interface areas in the range 1200-2000 A2 and only small conformation changes, so the rigid body approximation applies. It is less generally valid in protein-DNA complexes, which have interfaces covering 2200-3100 A2, large dissociation heat capacities, and affect both the conformation and the dynamics of their components.

The isomorphous structures of prethrombin2, hirugen-, and PPACK-thrombin: changes accompanying activation and exosite binding to thrombin

The X-ray crystal structure of prethrombin2 (pre2), the immediate inactive precursor of alpha-thrombin, has been determined at 2.0 A resolution complexed with hirugen. The structure has been refined to a final R-value of 0.169 using 14,211 observed reflections in the resolution range 8.0-2.0 A. A total of 202 water molecules have also been located in the structure. Comparison with the hirugen-thrombin complex showed that, apart from the flexible beginning and terminal regions of the molecule, there are 4 polypeptide segments in pre2 differing in conformation from the active enzyme (Pro 186-Asp 194, Gly 216-Gly 223, Gly 142-Pro 152, and the Arg 15-Ile 16 cleavage region). The formation of the Ile 16-Asp 194 ion pair and the specificity pocket are characteristic of serine protease activation with the conformation of the catalytic triad being conserved. With the determination of isomorphous structures of hirugen-thrombin and D-Phe-Pro-Arg chloromethyl ketone (PPACK)-thrombin, the changes that occur in the active site that affect the kinetics of chromogenic substrate hydrolysis on binding to the fibrinogen recognition exosite have been determined. The backbone of the Ala 190-Gly 197 segment in the active site has an average RMS difference of 0.55 A between the 2 structures (about 3.7 sigma compared to the bulk structure). This segment has 2 type II beta-bends, the first bend showing the largest shift due to hirugen binding. Another important feature was the 2 different conformations of the side chain of Glu 192. The side chain extends to solvent in hirugen-thrombin, which is compatible with the binding of substrates having an acidic residue in the P3 position (protein-C, thrombin platelet receptor). In PPACK-thrombin, the side chain of Asp 189 and the segment Arg 221A-Gly 223 move to provide space for the inhibitor, whereas in hirugen-thrombin, the Ala 190-Gly 197 movement expands the active site region. Although 8 water molecules are expelled from the active site with PPACK binding, the inhibitor complex is resolvated with 5 other water molecules.

Cold adaption of enzymes: structural comparison between salmon and bovine trypsins

The crystal structure of an anionic form of salmon trypsin has been determined at 1.82 A resolution. We report the first structure of a trypsin from a phoikilothermic organism in a detailed comparison to mammalian trypsins in order to look for structural rationalizations for the cold-adaption features of salmon trypsin. This form of salmon trypsin (ST II) comprises 222 residues, and is homologous to bovine trypsin (BT) in about 65% of the primary structure. The tertiary structures are similar, with an overall displacement in main chain atomic positions between salmon trypsin and various crystal structures of bovine trypsin of about 0.8 A. Intramolecular hydrogen bonds and hydrophobic interactions are compared and discussed in order to estimate possible differences in molecular flexibility which might explain the higher catalytic efficiency and lower thermostability of salmon trypsin compared to bovine trypsin. No overall differences in intramolecular interactions are detected between the two structures, but there are differences in certain regions of the structures which may explain some of the observed differences in physical properties. The distribution of charged residues is different in the two trypsins, and the impact this might have on substrate affinity has been discussed.

BPTI backbone variants and implications for inhibitory activity

Structural variants of BPTI were synthesized en route an enzymatic-chemical semisynthesis. The P1-P2 amide bond of the inhibitor molecule, which, as donor, contributes a hydrogen bond towards trypsin in the enzyme-inhibitor complex, was replaced by either a ketomethylene function or an ester bond yielding molecules with inhibitory activity. The two backbone-mutated BPTI derivatives showed increased dissociation constants of their respective trypsin complexes, obviously due to the lack of a single hydrogen-bond interaction in the enzyme-inhibitor complex.

Alteration of the specificity of ecotin, an E. coli serine proteinase inhibitor, by site directed mutagenesis

The gene of ecotin, an E. coli proteinase inhibitor, was cloned, and by site-directed mutagenesis the active site residue of the protein, Met84, was mutated to Lys, Arg and Leu. The recombinant wild-type and mutant inhibitors were overexpressed in E. coli, purified to homogeneity and their inhibitory effects on trypsin, chymotrypsin and elastase were compared. Of these serine proteinases trypsin is the most strongly inhibited by wild type ecotin and its mutants. According to our results the character of residue 84 of ecotin significantly but not dramatically modifies the specificity of the inhibitor.

Synthetic inhibitors of elastase

For more than two decades investigators around the world, in both academic and industrial institutions, have been developing inhibitors of human neutrophil elastase. A number of very elegant and insightful strategies have been reported. In the case of reversible peptidic inhibitors, this has resulted in the identification of some extremely potent compounds with dissociation constants in the 10(-11) M range. This is quite an accomplishment considering that these low molecular-weight inhibitors are only tri- and tetrapeptides. In the case of the heterocyclic-based inhibitors, the challenge of balancing the heterocycle's inherent reactivity and aqueous stability with the stability of the enzyme-inhibitor adduct has been meet by either using a latent, reactive functionality which is only activated within the enzyme, or by incorporating features which selectively obstruct deacylation but have little effect on the enzyme acylation step. The underlying goal of this research has been the identification of agents to treat diseases associated with HNE. Several animal models have been developed for evaluating the in vivo activity of elastase inhibitors, and compounds have been shown to be effective in all of these models by the intravenous, intratrachael or oral routes of administration. However, only a very small percentage of compounds have possessed all the necessary properties, including lack of toxicity, for progression into the clinic. The peptidyl TFMK ICI 200,880 (25-12) has many of the desired characteristics of a drug to treat the diseases associated with HNE: chemical stability, in vitro and in vivo activity, a long duration of action, and adequate metabolic stability. Currently ICI 200,880 is the only low molecular-weight HNE inhibitor known to be undergoing clinical trials, and may be the compound which finally demonstrates the clinical utility of a synthetic HNE inhibitor.

Intimations of K+ channel structure from a complete functional map of the molecular surface of charybdotoxin

The external vestibules of many K+ channels carry a high-affinity receptor for charybdotoxin, a peptide of known structure. Point mutations of a recombinant toxin identified the residues directly involved in the interaction with a Ca(2+)-activated K+ channel. The interaction surface is formed from 8 of the 37 residues and covers about 25% of the peptide's molecular surface. The shape of the toxin permits a deduced picture of the complementary receptor site in the external vestibule of the K+ channel.

Analysis of molecular recognition: steric electrostatic and hydrophobic complementarity

We discuss three important aspects of molecular recognition: steric, electrostatic and hydrophobic. Steric fit means that interacting atoms may not approach each other beyond their van der Waals radii and, simultaneously, crevices should be filled as densely as possible. Electrostatic fit requires the maximum ionic and polar (hydrogen bond or other) interaction between host and guest atoms while the hydrophobic fit corresponds to the association trend between apolar groups in an aqueous medium. Space-filling models, obtained by molecular graphics, illustrate steric complementarity while we use molecular electrostatic potentials (MEPs) and fields (MEFs) to investigate electrostatic and hydrophobic matching. Molecular regions with negative and positive MEPs attract and repel a positive probe charge, respectively, so we consider them as attracting each other. Furthermore we postulate that regions with MEFs of similar magnitude tend to associate more strongly than those with very different fields (similis simili gaudet principle). We apply the above rules to the study of complementarity in the trypsin-BPTI complex and in a crystalline association between styrene epoxide as guest and a camphor-based anthracene derivative as host. We discuss molecular similarity on the same footing as complementarity and give some examples on the application of the concept to the rationalization of relative strengths of trypsin inhibitors.

A structural model for the glutamate-specific endopeptidase from Streptomyces griseus that explains substrate specificity

We present a model for the three-dimensional structure of the glutamate-specific endopeptidase from Streptomyces griseus based on the crystal structures of other bacterial proteases of the trypsin family. For the first time a structural model is described which attempts to explain the basis of P1 glutamate specificity in serine proteases. Several important changes to the S1 pocket with respect to other members of the family of different specificity are described. Of particular interest is the presence of a histidine at position 213 and the substitution of Arg-138 by lysine. Other biochemical evidence concerning substrate preferences can be rationalized on the basis of the model.

Binding of bovine pancreatic trypsin inhibitor to heparin binding protein/CAP37/azurocidin. Interaction between a Kunitz-type inhibitor and a proteolytically inactive serine proteinase homologue

Heparin-binding protein (HBP; also known as CAP37 or azurocidin) is a member of the serine proteinase family. Evolution, however, has reverted this protein into a non-proteolytic form by mutation of two of the three residues of the active-site triad. Although proteolytically inactive, the human heparin-binding protein (hHBP) is still capable of binding bovine pancreatic trypsin inhibitor (BPTI). This was demonstrated by affinity chromatography to BPTI immobilized on a solid matrix and by studies on plasmin inhibition kinetics. hHBP competes with plasmin for BPTI and this effect on plasmin inhibition has been analyzed in terms of a kinetic model. A dissociation constant, Kd = 0.1 microM, was found for the interaction between BPTI and hHBP. The hHBP provides an example of a serine proteinase which has lost its catalytic function by reverting residues of the active center while still preserving its capability of specific interactions with Kunitz inhibitors. pHBP, the porcine counterpart to hHBP, on the other hand, was incapable of BPTI binding. The structural basis for the BPTI binding to the human protein and the species difference is discussed in terms of putative three-dimensional structures of the proteins derived by comparative molecular modelling methods.

Design of novel cationic ligands for the purification of trypsin-like proteases by affinity chromatography

A number of new cationic ligands have been designed and synthesized for the selective resolution and purification of the trypszin-like proteases. A series of ligands based on 4-[2'-methyl-4'-(2'',4''-dichloro-1'',3'',5''-triazin-6-ylamino ) phenylazo]benzamidine were able to bind to trypsin and the trypsin-like proteases, thrombin and urokinase, but bound pancreatic kallikrein only weakly. Ligands possessing a second cationic group (either 4-aminophenyltrimethylammonium or 4-aminobenzamidine) substituted onto the triazine ring displayed higher affinities than the parent compound for trypsin in solution but bound the enzyme weakly or not at all after immobilization. In contrast, these bis-cationic ligands bound pancreatic kallikrein in solution and following immobilization. The presence of the second cationic group was crucial, since its replacement by neutral or anionic groups led to loss of affinity for pancreatic kallikrein. One of the bis-cationic ligands was used to purify pancreatic kallikrein 9.5-fold from a crude pancreatic extract in 79% yield, to generate a product 99.9% free of contaminating trypsin activity.

The 0.25-nm X-ray structure of the Bowman-Birk-type inhibitor from mung bean in ternary complex with porcine trypsin

The structure of the Bowman-Birk-type inhibitor from mung bean Phaseolus aureus has been determined in ternary complex with porcine trypsin. The complex formed crystals of the trigonal space group P3(1)21 which diffracted to a resolution of 250 pm. Each of the two mung bean protease reactive sites is bound to trypsin according to the standard mechanism for serine proteinase inhibition. The binding loops thereby adopt the canonical conformation for the standard mechanism; however, the sub-van der Waals contact between the active-site serine O gamma (195) and the P1 carbonyl carbon of both loops is significantly smaller (210 pm) than hitherto observed, with continuous electron density connecting the two atoms. The inhibitor is formed by two double-stranded antiparallel beta-sheets, which are connected into a moderately twisted beta-sheet by a network of hydrogen bonds involving main-chain atoms and two water molecules. All contacts with neighbors in the crystal lattice occur between trypsin molecules. This apparently gives rise to an unusual form of disorder where the complexes pack in two orientations Ta:MaMb:Tb and Tb:MbMa:Ta (Ta, Tb = trypsin, Ma = mung bean loop I, Mb = mung bean loop II), such that the asymmetric unit consists of the ternary complex in two orientations, each with half occupancy. This is nearly equivalent to an asymmetric unit which has one trypsin molecule with full occupancy and one mung bean inhibitor with half occupancy and a crystallographic twofold symmetry axis through its center. Because of the approximate twofold symmetry of the inhibitor itself, however, the electron density was interpretable for most of the inhibitor (17 residues at the termini were not resolved) and shows evidence of its double orientation.

Bovine pancreatic trypsin inhibitor-trypsin complex as a detection system for recombinant proteins

Bovine pancreatic trypsin inhibitor (BPTI) binds to trypsin and anhydrotrypsin (an enzymatically inactive derivative of trypsin) with affinities of 6 x 10(-14) and 1.1 x 10(-13) M, respectively. We have taken advantage of the high affinity and specificity of this binding reaction to develop a protein tagging system in which biotinylated trypsin or biotinylated anhydrotrypsin is used as the reagent to detect recombinant fusion proteins into which BPTI has been inserted. Two proteins, opsin and growth hormone, were used as targets for insertional mutagenesis with BPTI. In each case, both domains of the fusion protein appear to be correctly folded. The fusion proteins can be specifically and efficiently detected by biotinylated trypsin or biotinylated anhydrotrypsin, as demonstrated by staining of transfected cells, protein blotting, affinity purification, and a mobility shift assay in SDS/polyacrylamide gels.

Identification of serine and histidine adducts in complexes of trypsin and trypsinogen with peptide and nonpeptide boronic acid inhibitors by 1H NMR spectroscopy

We have previously shown, in 15N NMR studies of the enzyme's active site histidine residue, that boronic acid inhibitors can form two distinct types of complexes with alpha-lytic protease. Inhibitors that are structural analogs of good alpha-lytic protease substrates form transition-state-like tetrahedral complexes with the active site serine whereas those that are not form complexes in which N epsilon 2 of the active site histidine is covalently bonded to the boron of the inhibitor. This study also demonstrated that the serine and histidine adduct complexes exhibit quite distinctive and characteristic low-field 1H NMR spectra [Bachovchin, W. W., Wong, W. Y. L., Farr-Jones, S., Shenvi, A. B., & Kettner, C. A. (1988) Biochemistry 27, 7689-7697]. Here we have used low-field 1H NMR diagnostically for a series of boronic acid inhibitor complexes of trypsin and trypsinogen. The results show that H-D-Val-Leu-boroArg and Ac-Gly-boroArg, analogs of good trypsin substrates, form transition-state-like serine adducts with trypsin, whereas the nonsubstrate analog inhibitors boric acid, methane boronic acid, butane boronic acid, and triethanolamine borate all form histidine adducts, thereby paralleling the previous results obtained with alpha-lytic protease. However, with trypsinogen, Ac-Gly-boroArg forms predominantly a histidine adduct while H-D-Val-Leu-boroArg forms both histidine and serine adducts, with the histidine adduct predominating below pH 8.0 and the serine adduct predominating above pH 8.0.(ABSTRACT TRUNCATED AT 250 WORDS)

Shape distributions of protein topography

A method is presented for measuring protein surface shape. It is an improvement of an earlier method that intersects a sphere with the solvent-excluded volume of a protein molecule. The new method, called a shape distribution, produces a more sophisticated description of the region of the sphere inside the protein than is provided by simply measuring the region's area or solid angle. This method is applied to the prediction of molecular complexes in three systems: the hemoglobin nonallosteric interface, trypsin and trypsin inhibitor, and heme and apomyoglobin. It does not uniquely predict the correct structure, even though the individual structures are taken from the experimentally determined complex structure. However, it does provide a list of several hundred predicted complexes, one of which is correct, and from which the correct complex might be extracted by a subsequent chemical filter.

Identification of a protein-binding surface by differential amide hydrogen-exchange measurements. Application to Bowman-Birk serine-protease inhibitor

The binding surface of soybean trypsin/chymotrypsin Bowman-Birk inhibitor in contact with alpha-chymotrypsin has been identified by measurement of the change in amide hydrogen-exchange rates between free and chymotrypsin-bound inhibitor. Exchange measurements were made for the enzyme-bound form of the inhibitor at pH 7.3, 25 degrees C using fast-flow affinity chromatography and direct measurement of exchange rates in the protein complex from one-dimensional and two-dimensional nuclear magnetic resonance spectra. The interface is characterized by a broad surface of contact involving residues 39 through 48 of the anti-chymotryptic domain beta-hairpin as well as residues 32, 33 and 37 in the anti-chymotryptic domain loop of the inhibitor. A number of residues in the anti-tryptic domain of the protein also have an altered exchange rate, suggesting that there are changes in the protein conformation upon binding to chymotrypsin. These changes in amide exchange behavior are discussed in light of a model of the complex based on the X-ray crystallographic structure of turkey ovomucoid inhibitor third domain bound to a alpha-chymotrypsin, and the structure of free Bowman-Birk inhibitor determined in solution by two-dimensional nuclear magnetic resonance spectroscopy. The chymotrypsin-binding loop of Bowman-Birk inhibitor in the model is remarkably similar to the binding loop conformation in crystal structures of enzyme-bound polypeptide chymotrypsin inhibitor-I from potatoes, turkey ovomucoid inhibitor third domain, and chymotrypsin inhibitor-II from barley seeds.

Design of novel affinity adsorbents for the purification of trypsin-like proteases

A number of ligands for the selective purification by affinity chromatography of the trypsin-like protease, porcine pancreatic kallikrein, were designed de novo by computer-aided molecular design. The ligands were designed to mimic the side-chains of a number of arginyl dipeptides and included a benzamidine moiety substituted on a triazine ring. The ligands displayed inhibitory activities against pancreatic kallikrein which mirrored the specificity constants of the dipeptides they were designed to mimic. The ligand with the highest affinity for the enzyme, an analogue of a Phe-Arg dipeptide, when immobilized to Sepharose CL-4B via a hexamethylene spacer arm, purified pancreatic kallikrein 110-fold in one step from a crude pancreatic acetone extract.