Structure-function analysis of a series of glucagon-like peptide-1 analogs.

We have used NMR in conjunction with measurements of functional bioactivity to define the receptor-binding structure of glucagon-like peptide-1 (GLP-1.) Identification of the important residues for binding was accomplished by the substitution of amino acids at sites that seemed likely, from an examination of the amino acid sequence and from previously published observations, to be important in the three-dimensional (3D) structure of the molecule. Identification of the receptor-bound conformation of GLP-1, because it is a flexible peptide, required constraint of the peptide backbone into a predetermined 3D structure. Constraint was achieved by the introduction of disulfide bonds and specific side chain-side chain cross-links. The biological relevance of the synthetic structure of each rigidified peptide was assessed by measurement of its ability to bind to the receptor present on RINm5F cells and to elicit a functional response, cyclic AMP production. NMR solution structures were obtained for the most biologically relevant of these analogs. The results of this study indicated that the residues necessary for the biological activity of GLP-1 occupy approximately three equally-spaced regions of the peptide 3D structure, at the corners of an equilateral triangle whose sides are, at a minimum estimate, 12-15A.

Inclusion complexation of ziprasidone mesylate with beta-cyclodextrin sulfobutyl ether.

Ziprasidone is an antipsychotic agent indicated primarily for the treatment of schizophrenia. An intramuscular dosage form of ziprasidone was developed using beta-cyclodextrin sulfobutyl ether (SBECD) to solubilize the drug by complexation. Inclusion complexation of ziprasidone mesylate (ZM) with SBECD was studied by circular dichroism (CD) spectroscopy, proton nuclear magnetic resonance (1H NMR) spectroscopy, Monte Carlo simulations, phase-solubility studies, and counterion titration. The results of the studies indicate that ZM, of which the counterion is not fully dissociated from the drug, forms a 1:1 inclusion complex with SBECD with the benzisothiazole group positioned in the cavity. A mathematical model was developed to calculate stability constants of inclusion complexes for the ion pair (Z+M-:SBECD) and the dissociated ionic form (Z+:SBECD) of ZM; the values were 7892 and 957 M(-1), respectively. The model also allowed the dissociation constants of noncomplexed and complexed ZM to be calculated; the value of the former is 8-fold greater than the value of the latter. These results indicate that the inclusion complex formation of the ion pair is favored over that of the dissociated ionic form of ZM, and that the dissociation of ZM is suppressed by inclusion complexation with SBECD.

Binding of a high-energy substrate conformer in antibody catalysis.

Enzymes can substantially increase the probability of a reaction by exploiting binding energy to preorganize their substrates into reactive conformations. Similar effects are likely to be important in a wide variety of designed catalysts, including catalytic antibodies. Transferred nuclear Overhauser effects have been used here to investigate how an antibody possessing chorismate mutase activity binds its flexible substrate molecule chorismate. The conversion of chorismate to prephenate by way of a Claisen rearrangement requires the substrate to adopt an energetically disfavored diaxial conformation in which the enolpyruvyl side chain is positioned over the six-membered ring. The antibody, which was elicited by a conformationally restricted transition state analog for this reaction, appears to bind this high-energy substrate conformer preferentially, as judged by diagnostic intramolecular transferred nuclear Overhauser effects. Inhibitor studies with the transition state analog confirm that preorganization takes place exclusively at the antibody active site. These results thus provide strong physical evidence for a direct relationship between the properties of a catalytic antibody and the structure of the transition state analog originally used to elicit the immune response.

Molecular structure of charybdotoxin: retraction.

Shortly after our paper of 3 August 1990 on the molecular structure of charybdotoxin (1) was published, two independent determinations of the structure of this molecule appeared (2) that were similar to each other and in strong disagreement with ours. We have obtained new data and find that some spectral features depend on solvent conditions, which explains some differences between our data and those of the other groups. More important, we conclude that we most probably misassigned an important sequence of amino acids, as suggested by Bontems et al.(3). Therefore, we withdraw our previously reported structure (1) and regret any inconvenience it may have caused. We thank F. Toma for sending us a copy of his paper before publication and for discussions.

Molecular structure of charybdotoxin, a pore-directed inhibitor of potassium ion channels.

The three-dimensional structure of charybdotoxin, a high-affinity peptide blocker of several potassium ion channels, was determined by two-dimensional nuclear magnetic resonance (2-D NMR) spectroscopy. Unambiguous NMR assignments of backbone and side chain hydrogens were made for all 37 amino acids. The structure was determined by distance geometry and refined by nuclear Overhauser and exchange spectroscopy back calculation. The peptide is built on a foundation of three antiparallel beta strands to which other parts of the sequence are attached by three disulfide bridges. The overall shape is roughly ellipsoidal, with axes of approximately 2.5 and 1.5 nanometers. Nine of the ten charged groups are located on one side of the ellipsoid, with seven of the eight positive residues lying in a stripe 2.5 nanometers in length. The other side displays three hydrophobic residues projecting prominently into aqueous solution. The structure rationalizes several mechanistic features of charybdotoxin block of the high-conductance Ca2(+)-activated K+ channel.

Catalysis of intermolecular oxygen atom transfer by nitrite dehydrogenase of Nitrobacter agilis.

Nitrobacter agilis, which contains a very active nitrite dehydrogenase, was studied in vivo under anaerobic conditions by the 15N NMR technique. When incubated with equimolar 15NO3- and unlabeled nitrite (or 15NO2- and unlabeled nitrate) the bacterium catalyzed an isotope exchange reaction at rates about 10% those observed in the nitrite oxidase assay. When incubated with 18O-labeled 15NO2- and 18O-labeled 15NO3-, the 18O was observed to exchange at similar rates from both species into water. Finally, when incubated with equimolar [18O]nitrate and 15NO2-, intermolecular 18O transfer was observed to result in formation of double labeled nitrate and nitrite at similar rates. 18O was transferred from nitrate to a 15N species or to water at approximately equal rates under the conditions of the experiments. It is argued that the enzyme responsible for these exchange reactions is nitrite dehydrogenase and not nitrate reductase. This work and the related experiments of DiSpirito and Hooper (DiSpirito, A.A., and Hooper, A.B. (1986) J. Biol. Chem. 261, 10534-10537) represent the first demonstrations of intermolecular oxygen atom transfer among oxotransferases. Mechanistic implications are discussed.