Molecular mechanics calculations on deaminooxytocin and on deamino-arginine-vasopressin and its analogues

Conformations of a model cyclic hexapeptide, CYIQNC: (1)H-NMR and molecular dynamics studies

Solution conformation of the cyclic hexapeptide sequence, [cyclo-S-Cys-Tyr-Ile-Gln-Asn-Cys-S] (CYIQNC) - a disulfide-linked fragment of a neurohypophyseal peptide hormone oxytocin (OT) - has been investigated by high-field one-dimensional (1D) and two-dimensional (2D) NMR spectroscopic methods and compared with the results obtained from computer simulation studies. (1)H-NMR results based on temperature dependence of amide proton chemical shifts and nuclear Overhauser effect indicate that peptide in solution populates different conformations, characterized by two fused β-turns. The segment Ile(3)-Gln(4)-Asn(5)-Cys(6) yields a preferred type-III β-turn at residues 4, 5 (HB, 3HN → 6CO), while the segment Cys(6), Cys(1)-Tyr(2)-Ile(3) exhibits inherently weaker, flexible β-turn either of type I/II'/III/half-turn at residues 1, 2 (HB, 6HN → 3CO). The computer simulation studies using a mixed protocol of distance geometry-simulated annealing followed by constrained minimization, restrained molecular dynamics, and energy minimization showed the possibility of existence of additional conformations with the hydrogen bonds, (a) 5HN → 3CO and (b) 2HN → 6CO. These results, therefore, indicate that the additional conformations obtained from both NMR and simulation studies can also be possible to the peptide. These additional conformations might have very small population in the solution and did not show their signatures in these conditions. These findings will be helpful in designing more analogs with modifications in the cyclic moiety of OT.

Conformational solution studies of (Sar7)desamino- and (MeAla7)desamino-vasopressin analogues using NMR spectroscopy

Solution structures of two analogues of vasopressin with an amino acid sequence of c[Mpa1-Tyr2-Phe3-Gln4-Asn5-Cys6]-Xaa7-Arg8-Gly9-NH2 (Xaa = Sar [I] or MeAla [II]) were established using ROE and the 3J(HNH alpha) couplings. Each of the peptides was found to exist in two stable isomers, pertaining to the cis or trans status of the Cys6-Xaa7 peptide bond, thus giving rise to four study objects. Two methods for studies of the conformational properties of the structures were compared. In the first, the algorithm consisted of three steps: (i) An Electrostatically Driven Monte-Carlo (EDMC) search for low-energy conformations. (ii) Simulations of the NOESY spectra and the vicinal couplings for these conformations. (iii) Determination of the statistical weights of the conformations with the ANALYZE package, so as to meet the best fit of the averaged NOE intensities and couplings to the experimental data. In the second method, the distance constraints and the torsion angles were used as the usual constraints in the Distance Geometry and Simulated Annealing algorithms. The flexibility of the pressin ring and the C-terminus was characterized by a large number of families of conformations. The presence of the beta-turn at position 4,5 was confirmed for all low energy structures found. The use of the EDMC method for the elaboration of the NMR data for small flexible peptides yielded an adequate description of their conformational diversity and is the method of choice for the analysis of their solution structures.

Fluorescence study of neurohypophyseal hormones and their analogues. Distance distributions in a series of arginine-vasopressin analogues

Analogues of arginine-vasopressin (AVP) in which substitution of the proline residue in position 7 (by either sarcosine or N-methylalanine) combined with replacement of the cysteine residue in position 1 were the subject of a fluorescence and molecular mechanics study. We obtained two groups of analogues: selective antidiuretic agonists (cysteine or beta-mercaptopropionic acid in position 1) and pressor and uterotonic antagonists (deaminopenicillamine or beta-mercapto-beta, beta-cyclopentamethylenepropionic acid in position 1). Using frequency-domain measurements of fluorescence resonance energy transfer (FRET) we estimated the distance distribution between the phenolic ring of Tyr2 and the disulphide bridge Cys1-Cys6. We also analyzed acrylamide quenching of tyrosyl fluorescence to determine the exposure of the tyrosyl ring to the solvent. Results from fluorescence experiments were compared with those from Monte Carlo simulation (ECEPP/3 force-field).

Exploration of the conformational space of oxytocin and arginine-vasopressin using the electrostatically driven Monte Carlo and molecular dynamics methods

Conformational analysis of the neurohypophyseal hormones oxytocin (OT) and arginine-vasopressin (AVP) has been carried out using two different computational approaches and three force fields, namely by the Electrostatically Driven Monte Carlo (EDMC) method, with the Empirical Conformational Energy Program for Peptides (ECEPP/3) force field or with the ECEPP/3 force field plus a hydration-shell model, and by simulated-annealing molecular dynamics with the Consistent Valence Force Field (CVFF). The low-energy conformations obtained for both hormones were classified using the minimal-tree clustering algorithm and characterized according to the locations of beta-turns in the cyclic moieties. Calculations with the CVFF force field located conformations with a beta-turn at residues 3 and 4 as the lowest energy ones both for OT and for AVP. In the ECEPP/3 force field the lowest energy conformation of OT contained a beta-turn at residues 2 and 3, conformations with this location of the turn being higher in energy for AVP. The latter difference can be attributed to the difference in the size of the side chain in position 3 of the sequences: the bulkier phenylalanine residue of AVP in combination with the bulky Tyr2 residue hinders the formation of a turn at residues 2 and 3. Conformations of OT and AVP with a turn at residues 3,4 were in the best agreement with the x-ray structures of deaminooxytocin and pressinoic acid (the cyclic moiety of vasopressin), respectively, and with the nmr-derived distance constraints. Generally, the low-energy conformations obtained with the hydration-shell model were in a better agreement with the experimental data than the conformations calculated in vacuo. It was found, however, that the obtained low-energy conformations do not satisfy all of the nmr-derived distance constraints and the nuclear Overhauser effect pattern observed in nmr studies can be fully explained only by assuming a dynamic equilibrium between conformations with beta-turns at residues 2,3, 3,4, and 4,5. The low-energy structures of OT with a beta-turn at residues 2,3 have the disulfide ring conformations close to the model proposed recently for a potent bicyclic antagonist of OT [M. D. Shenderovich et al. (1994) Polish Journal of Chemistry, Vol. 25, pp. 921-927], although the native hormone differs from the bicyclic analogue by the conformation of the C-terminal tripeptide. This finding confirms the hypothesis of different receptor-bound conformations of agonists and antagonists of OT.

Distance distributions from the tyrosyl to disulfide residues in the oxytocin and [Arg8]-vasopressin measured using frequency-domain fluorescence resonance energy transfer

We have examined the fluorescence intensity decays of oxytocin and [Arg8]-vasopressin resulting from the single tyrosyl residue in each peptide, and the intensity decay of the Asu1,6-analogues in which the disulfide bridge is substituted by a CH2-CH2 bridge. Viscosity-dependent steady state and intensity decay measurements indicated that fluorescence resonance energy transfer (FRET) from tyrosyl phenol to the disulfide bridge is responsible for the decrease in fluorescence relative to the Asu-analogues. The frequency-domain phase and modulation data for the tyrosyl donor were interpreted in terms of fluorescence resonance energy transfer (FRET) to the weakly absorbing disulfide bridge and a distribution of donor-to-acceptor distances. Energy transfer efficiencies were determined from both time-resolved and steady-state measurements. Fitting the frequency-domain phase and modulation data to a Gaussian distance distribution indicated that the average inter-chromophoric distance (Rav) is similar in both compounds, Rav = 7.94 A for oxytocin and Rav = 8.00 A for vasopressin. However, the width of the distance distribution is narrower for vasopression (hw = 2.80 A) than for oxytocin (hw = 3.58 A), which is consistent with restriction of the tyrosine phenol motion due to its stacking wih the Phe3 side chain of vasopressin. Finally, the recovered distance distribution functions are compared with histograms describing the distance between the chromophores during the course of long, in vacuo, molecular dynamics runs using the computer program CHARMm and the QUANTA 3.0 parameters.

Prediction of protein conformation on the basis of a search for compact structures: test on avian pancreatic polypeptide

Based on the concept that hydrophobic interactions cause a polypeptide chain to adopt a compact structure, a method is proposed to predict the structure of a protein. The procedure is carried out in four stages: (1) use of a virtual-bond united-residue approximation with the side chains represented by spheres to search conformational space extensively using specially designed interactions to lead to a collapsed structure, (2) conversion of the lowest-energy virtual-bond united-residue chain to one with a real polypeptide backbone, with optimization of the hydrogen-bond network among the backbone groups, (3) perturbation of the latter structure by the electrostatically driven Monte Carlo (EDMC) procedure, and (4) conversion of the spherical representation of the side chains to real groups and perturbation of the whole molecule by the EDMC procedure using the empirical conformational energy program for peptides (ECEPP/2) energy function plus hydration. Application of this procedure to the 36-residue avian pancreatic polypeptide led to a structure that resembled the one determined by X-ray crystallography; it had an alpha-helix starting at residue 13, with the N-terminal portion of the chain in an extended conformation packed against the alpha-helix. Similar structures with slightly higher energies, but looser packing, were also obtained.

Conformational analysis of [Cpp1, Sar7, Arg8] vasopressin by 1H-NMR spectroscopy and molecular mechanics calculations

A combined 1H-NMR and molecular mechanics study of [Cpp1, Sar7]AVP was performed in order to select the most probable conformations in DMSO solutions. The NMR constraints obtained were employed in the selection of starting conformations of the cyclic moiety of the analog. In particular, the diminished accessibility of the Asn5 NH proton to solvent and the close contact between Cpp1 and Cys6 C alpha H protons suggests a beta-turn conformation at the Phe3-Gln4 residues. Energy minimization was carried out both in the ECEPP/2 (rigid-valence geometry) and in the AMBER (flexible-valence geometry) force fields. Comparison of the experimental and calculated values of NMR characteristics has revealed that conformations containing type I, II, and III beta-turns at the Phe3-Gln4 residues are in reasonable agreement with the experimental data, with a dynamic equilibrium between the beta I (beta III) and beta II type structures of the cyclic part being the most probable. All of these conformations prefer the negative chirality of the disulfide bridge (theta 3 approximately -90 degrees). Five representative conformations were chosen for the acyclic tail: one with a beta I, one with a beta II'-turn at the Sar7-Arg8 residues, two extended-type conformations, and a conformation with a gamma-turn at Sar7. Because only high-energy extended conformations were in agreement with NMR data, it was concluded that the acyclic tail has considerable conformational flexibility in solution. The conformations obtained are discussed in terms of the structure-function relationship of the neurohypophyseal hormone analogs.

Conformation of [8-arginine]vasopressin and V1 antagonists in dimethyl sulfoxide solution derived from two-dimensional NMR spectroscopy and molecular dynamics simulation

Structural and dynamic properties of [8-arginine]vasopressin and a class of highly potent vasopressin V1 antagonists which contain 3-mercapto-3,3-cyclopentamethylene propionic acid (Mca) in position 1 of the vasopressin sequence have been determined. On the basis of two-dimensional NMR experiments in dimethyl sulfoxide solution, interproton distances were derived according to which model conformations were built and refined using molecular dynamics simulations. The conformation of vasopressin and the V1 antagonists differ mainly in the region of the mutated residue. The antagonistic property was found to be related to an inversed chirality of the disulfide bridge. In all investigated molecules, characteristic beta-turn structure elements were found for the backbone conformation of the endocyclic residues Tyr2-Asn5. For this portion of the peptide sequence, various conformational equilibria were detected which matched different time scales. For [Arg8]vasopressin, averaged NMR parameters were obtained which could be explained by rapid interconversion between different beta-turn geometries, whereas multiple slowly exchanging conformations were observed for the V1 antagonists. V1 antagonists containing sarcosine in position 7 exhibited multiple spectral patterns for the exocyclic part attributed to cis/trans isomerization. The X-ray structure of deamino-oxytocin [Wood, S. P., Tickle, I. J., Treharne, A. M., Pitts, J. E., Mascarenhas, Y., Li, J. Y., Husain, J., Cooper, S., Blundell, T. L., Hruby, V. J., Buku, A., Fischman, A. J. & Wyssbrod, H. R. (1986) Science 232, 633-636] was found to represent one sample of the conformational space covered by the multiple conformations found for [Mca1, Arg8]vasopressin.

Development and testing of protocols for computer-aided design of peptide drugs, using oxytocin

Considerable clinical interest in neuropeptides and peptide hormones has stimulated recent research and development of peptide-based drugs. This process differs from most classical drug discovery procedures because peptide molecules have considerable inherent flexibility. In the present paper, to identify lowest energy and metastable conformers for drug design, and to develop protocols for such studies, conformational search algorithms, incorporating empirical energy calculations, have been applied in the analysis of the peptide oxytocin. Energy minimization in torsion angle space was carried out from a variety of starting conformations, including published structures, in all-atom mode and all with distance constraints for disulphide bond formation. The energy-minimized conformations have been further optimized by a mapping method. Complementary simulations have been performed in united-atom mode and a model representing the effects of water using dummy sites has been developed and tested for this representation. Several of the preferred conformers together with de novo conformations have been used as starting points in molecular dynamics simulations; 28 low potential energy conformations were located at a temperature of 4 K. Conformations are analysed to identify hydrogen bonds, phi-psi angle distributions and the RMS values relative to the X-ray structure of deamino-oxytocin. The modelled structure of lowest energy in the molecular mechanics calculations was also that of least RMS deviation from the crystal structure; whilst structures of lower energy but larger deviation were identified by molecular dynamics techniques. A metastable structure has been identified which satisfies existing criteria for the "active form", and this model is tested by a theoretical residue-substitution technique, to provide clues on the agonist/antagonist relationship at the atomic level.

A theoretical study of glucosamine synthase. Part I. Molecular mechanics calculations on substrate binding

Glucosamine synthase transfers the gamma-amino group of glutamine to fructose, producing 1-glucosamine which is the key constituent of bacterial and fungal cell walls. In this study, model calculations were performed on substrate binding to the enzyme active site. Two models of the active site of glucosamine synthase were proposed, which assume two different sequences of aminoacids, Cys-Gly-Ile and Cys-Ala-Cys, the first one being the N-terminal sequence of the Escherichia coli enzyme. Several initial geometries were assumed for these tripeptides, the energy was then optimized by means of molecular mechanics. It has been found that the structure which is both energy optimal and satisfies the assumed cysteine sulphur arrangement consists of combinations of C7eq and C7ax conformations of single residues. Molecular mechanics calculations were then performed on glutamine and D-fructose-6-phosphate, which are the substrates of the enzymatic catalysis, and on their complex with the enzyme glutamine-binding site. The spatial configuration of the compounds under study, which is optimal as far as the reaction path is concerned, also turned out to be an energy minimum.

Theoretical studies of the mechanism of the action of the neurohypophyseal hormones. I. Molecular electrostatic potential (MEP) and molecular electrostatic field (MEF) maps of some vasopressin analogues

Continuing our theoretical studies of the oxytocin and vasopressin analogues, we have analysed the molecular electrostatic potential (MEP) and the norm of the molecular electrostatic field (MEF) of [1-beta-mercaptopropionic acid]-arginine-vasopressin ([ Mpa1]-AVP), [1-(beta-mercapto-beta,beta-cyclopentamethylene)propionic acid]-arginine-vasopressin ([Cpp']-AVP), and [1-thiosalicylic acid]-arginine-vasopressin ([Ths']-AVP) whose low-energy conformations were calculated in our previous work. These compounds are known from experiment to exhibit different biological activity. The scalar fields mentioned determine the energy of interaction with either charged (MEP) or polar (MEF) species, the energy being in the second case either optimal or Boltzmann-averaged over all the possible orientations of the dipole moment versus the electrostatic field. The electrostatic interactions slowly vanish with distance and can therefore be considered to be the factor determining the molecular shape at greater distances, which can help in both predicting the interactions with the receptor at the stage of remote recognition and in finding the preferred directions of solvation by a polar solvent. In the analysis of the fields three techniques have been used: (i) the construction of maps in certain planes; (ii) the construction of maps on spheres centered in the charge center of the molecule under study and of poles chosen according to the main axes of the quadrupole moment; and (iii) the construction of surfaces corresponding to a given value of potential. The results obtained show that the shapes of both MEP and MEF are similar in the case of [Mpa1]-AVP and [Cpp1]-AVP (biologically active), while some differences emerge when comparing these compounds with [Ths1]-AVP (inactive). It has also been found that both MEP and MEF depend even more strongly on conformation.