X-ray and NMR characterization of covalent complexes of trypsin, borate, and alcohols

An understanding of the physiological and toxicological properties of borate and the utilization of boronic acids in drug development require a basic understanding of borate-enzyme chemistry. We report here the extension of our recent NMR studies indicating the formation of a ternary borate-alcohol-trypsin complex. Crystallographic and solution state NMR studies of porcine trypsin were performed in the presence of borate and either of three alcohols designed to bind to the S1 affinity subsite: 4-aminobutanol, guanidine-3-propanol, and 4-hydroxymethylbenzamidine. Quaternary complexes of trypsin, borate, S1-binding alcohol, and ethylene glycol (a cryoprotectant), as well as a ternary trypsin, borate, and ethylene glycol complex have been observed in the crystalline state. Borate forms ester bonds to Ser195, ethylene glycol (two bonds), and the S1-binding alcohol (if present). Spectra from (1)H and (11)B NMR studies confirm that these complexes also exist in solution and also provide evidence for the formation of ternary trypsin, borate, and S1-subsite alcohol complexes which are not observed in the crystals using our experimental protocols. Analysis of eight crystal structures indicates that formation of an active site borate complex is in all cases accompanied by a significant (approximately 4%) increase in the b-axis dimension of the unit cell. Presumably, our inability to observe the ternary complexes in the crystalline state arises from the lower stability of these complexes and consequent inability to overcome the constraints imposed by the lattice contacts. A mechanism for the coupling of the lattice contacts with the active site that involves a conformational rearrangement of Gln192 is suggested. The structures presented here represent the first crystallographic demonstration of covalent binding of an enzyme by borate.     

13C NMR studies of the binding of soybean trypsin inhibitor to trypsin.

NMR studies of the complex between trypsin and soybean trypsin inhibitor with 1-¹³C-arginine and modified inhibitor with 1-¹³C-lysine show that these complexes involve almost exclusively non-covalent binding of the inhibitor to the enzyme for trypsin/¹³C-Lys-inhibitor at pH 6.5 and 8.1 and for trypsin/¹³C-Arg-inhibitor at pH 5.0. At pH 7.1 for trypsin/¹³C-Arg-inhibitor both non-covalent and acyl enzyme forms are observed. Under no conditions did we observe evidence for a tetrahedral adduct between enzyme and inhibitor.     

NMR detection of adventitious xylose binding to the quorum-sensing protein SdiA of Escherichia coli

During the solution structure determination of the Escherichia coli quorum-sensing protein SdiA in the presence of N-octanoyl-l-homoserine lactone (HSL), NMR signals were detected in (13)C-filter-(13)C-filter spectra for the bound HSL molecule. An additional set of coupled signals, independent of those of HSL, were also detected, indicating the presence of another unlabeled molecule, also bound to the labeled SdiA. Analysis of the NMR spectrum of this ligand and of the mass spectrum of the dissociated components indicates that the ligand is most likely xylose. Further analysis of xylose-bound SdiA defines a site close to the C terminus, remote from the HSL binding site. These observations provide an example of the sensitivity of high-resolution NMR experiments and their ability to detect, identify, and map the adventitious binding of a small organic molecule to a protein.     

Metal binding sites in proteins: identification and characterization by paramagnetic NMR relaxation

A method is presented that allows the identification and quantitative characterization of metal binding sites in proteins using paramagnetic nuclear magnetic resonance spectroscopy. The method relies on the nonselective longitudinal relaxation rates of the amide protons and their dependence on the paramagnetic metal ion concentration and the pH, and on the three-dimensional structure of the protein. The method is demonstrated using Escherichia coli thioredoxin as a model protein and Ni(2+) as the paramagnetic metal ion. Through a least-squares analysis of the relaxation rates, it is found that Ni(2+) binds to a series of specific sites on the surface of thioredoxin. The strongest binding site is found near the N-terminus of the protein, where the metal ion is coordinated to the free NH(2) group of the N-terminal serine residue and the side chain carboxylate group of the aspartic acid residue in position 2. In addition, Ni(2+) binds specifically but more weakly to the surface-exposed side chain carboxylate groups of residues D10, D20, D47, and E85.     

X-ray crystallographic and high-resolution NMR spectroscopy characterization of 4,6-di-O-acetyl-2,3-dideoxy-alpha-D-erythro-hex-2-enopyranosyl sulfamide

Single crystal X-ray diffraction and high-resolution ¹H and ¹³C NMR spectral data for 4,6-di-O-acetyl-2,3-dideoxy-α-d-erythro-hex-2-enopyranosyl sulfamide, a selective inhibitor of carbonic anhydrase isozyme IX, are reported. The ⁰H₅ was found to be the preferred form for this glycosyl sulfamide, both in the crystal lattice and in solution.     

A soybean trypsin inhibitor. Crystallization and x-ray crystallographic study

Five trypsin and alpha-chymotrypsin inhibitors which have low molecular weights (ranging from 6800 to 8600) and are present in soybean seeds of the Tracy variety have been isolated and purified, and single crystals which give x-ray diffraction data beyond 3-A spacings have been obtained from one of them. The trypsin inhibitor crystallizes in a monoclinic unit cell of symmetry P2(1) and dimensions a = 25.919(7) A, b = 43.23(1) A, c = 19.905(5) A, and beta = 103.63(2) degrees. The assymmetric unit contains 1 molecule of molecular weight 6800. The crystal, which has been found to be unusually stable to x-radiation, has solvent content of approximately 26% by volume.     

Structural interpretation of lanthanide binding to the basic pancreatic trypsin inhibitor by 1H NMR at 360 MHz.

The weak binding of lanthanides to the five carboxyl groups of the basic pancreatic trypsin inhibitor (hereafter termed "the inhibitor"), has been investigated in detail using high resolution 1H NMR at 360 MHz. Lanthanides bind to the C-terminus with an apparent binding constant of 30 M-1, and thus competitively inhibit the formation of a salt-bridge between the C-terminus and the N-terminus, Lanthanides bind also to the side chain carboxyl groups of Asp 3, Glu 7, Glu 49 and Asp 50, with binding constants of 10--30 M-1. With the use of lanthanides individual resonance assignments for Phe 4 and Phe 45 were obtained in the 1H NMR spectrum of the inhibitor, and for several spin systems previous identifications were independently confirmed. The present experiments also provide a nice illustration for the use of shift reagents to improve the resolution in 1H NMR spectra of proteins. The exchange broadening for Tyr 35 and Phe 45 over the temperature range 4--72 degrees C could thus be observed for almost all the components of these aromatic spin systems and new details on the dynamic properties were obtained also for other aromatic residues.     

NMR and crystallographic structures of the FK506 binding domain of human malarial parasite Plasmodium vivax FKBP35

The emergence of drug‐resistant malaria parasites is the major threat to effective malaria control, prompting a search for novel compounds with mechanisms of action that are different from the traditionally used drugs. The immunosuppressive drug FK506 shows an antimalarial activity. The mechanism of the drug action involves the molecular interaction with the parasite target proteins PfFKBP35 and PvFKBP35, which are novel FK506 binding protein family (FKBP) members from Plasmodium falciparum and Plasmodium vivax, respectively. Currently, molecular mechanisms of the FKBP family proteins in the parasites still remain elusive. To understand their functions, here we have determined the structures of the FK506 binding domain of Plasmodium vivax (PvFKBD) in unliganded form by NMR spectroscopy and in complex with FK506 by X‐ray crystallography. We found out that PvFKBP35 exhibits a canonical FKBD fold and shares kinetic profiles similar to those of PfFKBP35, the homologous protein in P. falciparum, indicating that the parasite FKBP family members play similar biological roles in their life cycles. Despite the similarity, differences were observed in the ligand binding modes between PvFKBD and HsFKBP12, a human FKBP homolog, which could provide insightful information into designing selective antimalarial drug against the parasites.     

Diastereotopic covalent binding of the natural inhibitor leupeptin to trypsin: detection of two interconverting hemiacetals by solution and solid-state NMR spectroscopy

The naturally occurring peptidyl protease inhibitor leupeptin (N-acetyl-L-leucyl-L-leucyl-L-argininal) has been prepared labeled with 13C at the argininal carbonyl. 13C chemical shift data for the trypsin-leupeptin inhibitor complex in the pH range 3.0-7.6 reveal the presence of two pH-dependent covalent complexes, suggestive of two interconverting diastereomers at the new asymmetric tetrahedral center created by covalent addition of Ser195 to either side of the 13C-enriched aldehyde of the inhibitor. At pH 7 two signals are observable at delta 98.8 and delta 97.2 (84:16 ratio), while at pH 3.0 the latter signal predominates. In the selective proton 13C-edited NOE spectrum of the major diastereomer at pH 7.4, a strong NOE is observed between the hemiacetal proton of the inhibitor and the C2 proton of His57 of the enzyme, thus defining the stereochemistry of the high pH complex to the S configuration in which the hemiacetal oxygen resides in the oxyanion hole. pH titration studies further indicate that the 13C chemical shift of the S diastereomer follows a titration curve with a pKa of 4.69, the magnitude of which is consistent with direct titration of the hemiacetal oxygen. Similar pH-dependent chemical shifts were obtained by using CPMAS 13C NMR, providing evidence for the existence of the same diastereomeric equilibrium in the solid state.     

Preliminary crystallographic characterization of ricin agglutinin

The quaternary structure of ricin agglutinin (RCA) has been determined by x-ray crystallography. The refined structure of ricin proved to be a successful search model using the molecular replacement method of phase determination. RCA forms an elongated molecule of dimensions 120 Å × 60 Å × 40 Å with two A chains at the center and a B chain at each end. The A chains are covalently associated via a disulfide bridge between Cys 156 of both chains. Additional contacts at residues 114–115 stabilize the dimer interface. The covalent association of RCA A chains was confirmed by gel filtration under reducing and nonreducing conditions. Proteins 28:586–589, 1997. © 1997 Wiley-Liss, Inc.     

NMR characterization of substrate binding in the phthalate dioxygenase system.

The paramagnetic enhancements in the NMR relaxation rates for the fluorine in fluorophthalates have been used to determine the position of the phthalate with respect to the mononuclear metal ion in native and metal-substituted derivatives of phthalate dioxygenase (PDO). These studies show directly that the substrate interacts with the mononuclear metal of PDO and provide the first structural characterization of this interaction. With a molecular mass of 200 kDa, PDO is one of the largest proteins studied to date by paramagnetic NMR. Two paramagnetically broadened (19)F lines were observed for monofluorophthalates bound to CoPDO. This demonstrates that fluorophthalate binds to PDO with a handedness, i.e., with the fluorine label facing to the "right" or to the "left", relative to the hyperfine tensor of the Co(II). The relative affinities of the two orientations are slightly different, with a 2-fold and 5-fold excess of the preferred orientation for 4-fluorophthalate and 3-fluorophthalate, respectively. The longitudinal relaxation rate (T(1)) and transverse relaxation rate (T(2)) data give mutually consistent fluorine to cobalt distances. These results are consistent with approximate bilateral symmetry, with the Co to 3-fluorophthalate distances ( approximately 5.5 A) approximately 25% longer than the Co to 4-fluorophthalate distances ( approximately 4. 5 A). A detailed geometric model is derived from these data. This structural characterization of the mononuclear site provides a framework to develop hypotheses for the mechanism of oxygenation by the Fe(II)-containing aromatic dioxygenases.     

Homotropic cooperative binding of organic solvent vapors by solid trypsin

Homotropic cooperative binding was observed at vapor sorption of organic solvents (acetonitrile, propionitrile, ethanol, 1-propanol, 2-propanol, nitroethane) by dried solid trypsin from porcine pancreas (0.05 g H2O/g protein). The vapor sorption isotherms were obtained by the static method of gas chromatographic headspace analysis at 298 K for 'vapor solvent+solid trypsin' systems in the absence of the liquid phase. All isotherms have a sigmoidal shape with significant sorbate uptake only above the threshold of sorbate thermodynamic activity. On the sorption isotherms of non-hydroxylic sorbates the saturation of trypsin by organic solvent was observed above the sorbate threshold activity. The formation of inclusion compounds with phase transition between solvent-free and solvent-saturated trypsin is supposed. Approximation of obtained isotherms by the Hill equation gives the inclusion stoichiometry S, inclusion free energy, and the Hill constant N of clathrates. The inclusion stoichiometry S depends significantly on the size and shape of sorbate molecules and changes from S=31 mol of sorbate per mol of trypsin for ethanol to S=6 for nitroethane. The inclusion free energies determined for the standard states of pure liquid sorbate and infinitely dilute solution in toluene are in the range from -0.5 to -1.2 kJ/mol and from -3.1 to -8.1 kJ/mol, respectively, per 1 mol of sorbate. The Hill constants are relatively high: from N=5.6 for 1-propanol to N approximately equal to 10(3) for nitroethane. The implication of the obtained results for the interpretation of solvent effects on the enzyme activity and stability in low-water medium is discussed.     

The perturbation of thrombin binding to human platelets by trypsin and neuraminidase

The initial step in the interaction of thrombin with human platelets in binding of the enzyme to the platelet surface. The effects of digestion of isolated platelets with trypsin and neuraminidase on aggregation, release of serotonin and binding of thrombin have been examined.Trypsin is a powerful inducer of platelet aggregation as well as the release reaction. The aggregation effect of trypsin may be blocked with disodium ehtylenediaminetatraacetate (EDTA). Further, in the presence of EDTA, trypsin-induced release of [¹⁴C]serotonin is 15–20% lower compared to controls and the initial lag period is prolonged. Conditions were developed under which trypsin did neither aggregate nor release serotonin from platelets. Even under these conditions, trypsin caused a profound loss in the thrombin binding capacity of platelets. Thus, the trypsin-induced fall in the thrombin binding capacity and the platelet response are dissociated. This loss in the thrombin binding by trypsin is due to a lower number of binding sites available on the platelet surface and is not due to an altered affinity.Neuraminidase did not induce platelet aggregation or the release reaction. The ability of platelets to bind thrombin was also unimpaired by prior digestion with neuraminidase. Thus, the sialic acid at the platelet surface is not essential in the function of thrombin recognition by the receptor. This moiety may nontheless be a constituent of a glycoprotein which might act as the thrombin receptor.     

Anion binding by human lactoferrin: results from crystallographic and physicochemical studies.

The anion-binding properties of lactoferrin (Lf), with Fe3+ or Cu2+ as the associated metal ion, have been investigated by physicochemical and crystallographic techniques. These highlight differences between the two sites and in the anion-binding behavior when different metals are bound. Carbonate, oxalate, and hybrid carbonate-oxalate complexes have been prepared and their characteristic electronic and EPR spectra recorded. Oxalate can displace carbonate from either one or both anion sites of Cu2(CO3)2Lf, depending on the oxalate concentration, but no such displacement occurs for Fe2(CO3)2Lf. Addition of oxalate and the appropriate metal ion to apoLf under carbonate-free conditions gives dioxalate complexes with both Fe3+ and Cu2+, except when traces of EDTA remain associated with the protein, when hybrid complexes M2(CO3)(C2O4)Lf can result. The anion sites in the crystal structures of Fe2(CO3)2Lf, Cu2-(CO3)2Lf, and Cu2(CO3)(C2O4)Lf, refined at 2.2, 2.1, and 2.2 A, respectively, have been compared. In every case, the anion is hydrogen bonded to the N-terminus of helix 5, an associated arginine side chain, and a nearby threonine side chain. The carbonate ion binds in bidentate fashion to the metal, except in the N-lobe site of dicupric lactoferrin, where it is monodentate; the difference arises from slight movement of the metal ion. The hybrid complex shows that the oxalate ion binds preferentially in the C-lobe site, in 1,2-bidentate mode, but with the displacement of several nearby side chains. These observations lead to a generalized model for synergistic anion binding by transferrins.     

Purification and characterization of trypsin produced by gut bacteria from Anticarsia gemmatalis

Purification of active trypsin in the digestive process of insects is essential for the development of potent protease inhibitors (PIs) as an emerging pest control technology and research into insect adaptations to dietary PIs. An important aspect is the presence of proteolytic microorganisms, which contribute to host nutrition. Here, we purified trypsins produced by bacteria Bacillus cereus, Enterococcus mundtii, Enterococcus gallinarum, and Staphylococcus xylosus isolated from the midgut of Anticarsia gemmatalis. The trypsins had a molecular mass of approximately 25 kDa. The enzymes showed increased activity at 40°C, and they were active at pH values 7.5–10. Aprotinin, bis‐benzamidine, and soybean Kunitz inhibitor (SKTI) significantly inhibited trypsin activity. The l ‐1‐tosyl‐amido‐2‐phenylethylchloromethyl ketone (TPCK), pepstatin A, E‐64, ethylenediamine tetraacetic acid, and calcium ions did not affect the enzyme activity at the concentrations tested. We infer the purified trypsins do not require calcium ions, by which they differ from the trypsins of other microorganisms and the soluble and insoluble trypsins characterized from A. gemmatalis. These data suggest the existence of different isoforms of trypsin in the velvetbean caterpillar midguts.