The role of backbone conformation in deltorphin II binding: a QSAR study of new analogues modified in the 5-, 6-positions of the address domain.

The delta selectivity of the opioid heptapeptides deltorphin I and II has been attributed to the C-terminal 'address' domain, the hydrophobic Val(5)-Val(6) residues apparently playing a topographical role. We now report the synthesis, opioid binding affinities, and a QSAR study of a series of peptides in which one of the valine side chains was altered. QSAR analyses included previously published models for a binding pocket interaction and an optimum size (Schullery, S.; Mohammedshah, T.; Makhlouf, H.; Marks, E.; Wilenkin, B.; Escobar, S.; Mousigian, C.; Heyl, D. Bioorg. Med. Chem. 1997, 5, 2221), and a new approach for backbone conformational effects using Langevin dynamics simulation (PM3 semi-empirical force field) of an isolated peptide fragment containing the side chain and flanking peptide bonds. No evidence is found of binding pocket interactions or optimum size for either the position-5 or -6 side chain. Rather, delta binding is generally disfavored while mu binding is either unaffected (position-5) or favored (position-6) by larger side chains. The dynamics results provide evidence of similar 'local' conformation roles for the positions 5 and 6 side chains. Specifically, delta binding is favored by side chains that maximize the extension of the backbone, measured as the through-space distance between peptide fragment ends, the angle between lines connecting the alpha-carbon with fragment ends, or the difference between the psi and phi peptide angles.

Binding to delta and mu opioid receptors by deltorphin I/II analogues modified at the Phe3 and Asp4/Glu4 side chains: a report of 32 new analogues and a QSAR study.

The synthesis and binding affinities of 32 X3Gly4 dual-substitution analogues of the natural opioid heptapeptides deltorphin I and II are reported. A multiple regression QSAR analysis was performed using those results along with literature data for the X3Asp4 and Phe3X4 side chain analogues. Fitting to a three-term potential well model with hydrophobic and van der Waals attraction terms and a steric repulsion term indicates that the delta and mu receptor sites for binding the residue three side chain are similar, and that the binding interaction is primarily van der Waals and secondarily hydrophobic. Further analysis indicates that both sites are more constrained with respect to side chain length than width or thickness, and the mu site appears to be somewhat larger. A binding model consistent with these findings pictures the native third residues Phe ring laying on a step notched out of the receptor surface, pointing toward the back (riser) of the step, and sandwiched between the receptor and ligand. However, the binding sites for the residue four side chains are quite different on delta and mu receptors. Binding to the delta site appears to involve both electrostatic attraction (probably to a partial positive charge) and van der Waals attraction, but not necessarily hydrogen bonding, and more constraint with respect to side chain length than width or thickness. In contrast, there is no evidence for any kind of binding attraction between the side chain of residue four and the mu site, which acts more as steric repulsion site, as though the space that is a pocket on the delta receptor is filled in on the mu receptor. A regression model based only on steric repulsion by van der Waals bulk and/or the effective bulk of a hydration layer accounts for over 80% of the residue four related variation in mu affinity.

Effect of membrane additives on vesicle fusion.

A large variety of alkyl derivatives were found to either slow or block the low-temperature induced fusion of dipalmitoylphosphatidylcholine small unilamellar vesicles (DPPC SUV) when incorporated into the SUV bilayer at five mol%. Only corn oil was fusogenic.

Dipalmitoylphosphatidylcholine-palmitic acid phase diagram studied by 13C nuclear magnetic resonance.

The phase diagram of dipalmitoylphosphatidylcholine (DPPC) and palmitic acid mixtures in excess D2O was studied by 13C-NMR. Phase boundaries were determined from plots of apparent spin-spin relaxation time T2 (for both choline methyl and fatty acid chain carbons) versus temperature. A peritectic transition in the 1-10 mol% region, whose existence has been theoretically inferred from the Gibbs phase rule but which was undetectable by differential thermal analysis (DTA) (S.E. Schullery et al. Biochemistry, 20 (1981) 6818-6824), was located by NMR at 41.6 degrees C. A second, nearby peritectic line at 44 degrees C, which had been shown by DTA to extend from about 3-25 mol% palmitic acid, was seen by NMR only above 10 mol%. The palmitic acid/DPPC complex (2:1), with a sharp melting point at 64 degrees C, reported in earlier studies, was also seen by NMR. A phase diagram including both NMR and DTA results is presented. Important general conclusions from this study are: (i) NMR and scanning thermal analysis are complementary techniques for phase studies; each can see transitions that are invisible to the other. (ii) The case for the applicability of the Gibbs phase rule to lipid bilayer systems has been strengthened by the observance of two predicted, close-spaced boundaries. (iii) Low concentrations of fatty acids and related molecules can not be assumed to disperse as simple ideal solutes in the bilayer matrix.

Differential thermal analysis of dipalmitoylphosphatidylcholine--fatty acid mixtures.

Mixtures of dipalmitoylphosphatidylcholine (DPPC) with palmitic, stearic, and myristic acids and the sodium salts of these acids were analyzed by differential thermal analysis (DTA) over a wide range of lipid compositions, all in excess water. All three fatty acids raise the liquid-crystal phase transition temperature and form sharp-melting complexes, with 1:2 DPPC--fatty acid stoichiometry observed for palmitic and stearic acids and suggested for myristic acid. Phase diagrams of the peritectic type, indicating nonideal mixing, was fitted to the DPPC--palmitic acid and DPPC--stearic acid data. In contrast, DPPC forms nearly ideal mixtures with the putative DPPC--myristic acid complex. At levels of only a few mole percent, both sodium stearate and myristate remove the pretransition and main transition and produce new peaks at approximately 30 and approximately 48 degrees C; the relative areas of the new peaks were unreproducible for the DPPC--myristate system. Sodium palmitate is the least disruptive of any of the sodium soaps or fatty acids; up to 80 mol % palmitate, the transition is lowered 3 degrees C and approximately doubled in width. The pretransition is detectable up to 36 mol %, and the main transition persists up to 88 mol % palmitate. The apparent pK of palmitic acid (12 mol %) in DPPC bilayers was determined to be 10.2 by direct pH measurement of ternary DPPC mixtures with known palmitic acid/sodium palmitate ratios; the intrinsic pK is estimated to be less than or approximately 8.5.

Fusion of dipalmitoylphosphatidylcholine vesicles.

Small unilamellar dipalmitoylphosphatidylcholine vesicles formed by sonication are shown to fuse spontaneously below the phase transition temperature. The ultimate fusion products are unilamellar vesicles about 700 A in diameter, which are stable and provide an intact ionic permeation barrier either above or below the phase transition. The fused vesicles have been characterized by gel chromatography, trapped volume, 31P nuclear magnetic resonance, and negative stain and freeze-fracture electron microscopy.

Gel filtration of egg phosphatidylcholine vesicles.

The abilities of Sepharose 2B (Pharmacia), Controlled Pore Glass (Electro-Nucleonics) and Bio-Gel A150m (Bio-Rad) to purify small unilamellar vesicles prepared by sonication and the ethanol-injection methods were compared. The Bio-Gel causes complete aggregation of the sonicated vesicles and partial aggregation of the ethanol-injection vesicles. Both Sepharose and Controlled Pore Glass are acceptable for purifying vesicles from multilamellar liposomes; however, neither will separate the vesicles from sonication by-products which might be formed.

Binding of uranyl to phosphatidylcholine liposomes. Liposome aggregation effect on surface area.

The binding of uranyl ion, UO2+2, to egg phosphatidylcholine liposomes was studied as a potential method for the determination of liposome surface areas. Unbound uranyl was determined spectrophotometrically as the Arsenazo III complex with centrifuge supernatant. There is an apparent positive cooperativity in uranyl binding at phosphatidylcholine concentrations above approx. 0.1 mM. The binding capacity per mol increases upon liposome dilution. The data are consistent with liposomes existing in a highly aggregated state. The binding constant in the limit of low concentration of bound uranyl was 9+/-3)-10(6) M-1 in 0.1 M NaCl, pH 4.1. At saturation about four uranyl ions are bound per 100 phosphatidylcholine molecules. Relative surface areas of different dispersions may be calculated from intercepts of extrapolated binding isotherms, and absolute surface areas may be calculated if a value for the uranyl-phosphatidylcholine stoichiometry is assumed.

Permeability of iodide in multilamellar liposomes modeled by two compartments and a reservoir.

A previously published rate law for the diffusion of iodide from multilamellar egg phosphatidylcholine liposomes (Schullery, S.E. (1975) Chem. Phys. Lipids 14, 49-58) is fitted to the relatively simple mathematical model of two compartments in series with a reservoir. All of the inner liposome compartments are assumed to behave as effectively one compartment in series with the liposome's outermost compartment. Based on this model, reasonable values are calculated for the fraction of the total solution trapped by liposomes which is in the outermost liposome compartment, 17%, and the permeability coefficient of iodide against isotonic, mixed iodide-chloride solution, 2-10(-9) cm/s.

Studies on phosphatidylcholine model membranes. I. Iodine permeability measurement by specific ion electrode.

An iodide specific ion electrode was used to measure iodide released from egg phosphatidylcholine liposomes after passage through an ion exchange column. The permeation process was shown to be a sum of two separate first order processes. The method permits linear initial rates of iodide release to be determined for the first 5 min of permeation. Iodide permeability in the absence of Tris buffer was found to be a decreasing function of pH. In the presence of Tris, iodide permeability went through a minimum ofrom ph 7.3-8.5. the permeability was found to decrease when cholesterol was incorporated into the liposomes.