Interaction of a highly potent dimeric enkephalin analog, biphalin, with model membranes.

Biphalin, (Tyr-D-Ala-Gly-Phe-NH)2, is a highly potent dimeric analog of enkephalin. Its analgesic efficacy is due in part to its ability to permeate the blood-brain barrier. To aid in understanding the mechanism of the transmembrane movement we determined and analyzed the permeability and partition coefficients of biphalin and a series of analogues where F, Cl, I, NO2, or NH2 were placed in the para position of the aromatic rings of Phe4,4'. Liposomes composed of neutral phospholipids and cholesterol were used as the model membrane. The overall good correlation between permeability and water-membrane partition coefficients suggests that the movement of biphalins across the model membrane is controlled by diffusion and depends on the water-membrane partition coefficient. To explain the observed correlation between permeability and the electron withdrawing/donating character of the substituents in the phenylalanine ring, we examined various folding patterns of Leu-enkephalin, an endogenous pentapeptide that exhibits affinities toward the same classes of opioid receptors (delta and mu). The observed permeabilities and partition coefficients of biphalin and analogues, as well as the tyrosine side chain accessibility, are consistent with the presence of the type of folding where the tyrosine and phenylalanine side chains are in a close contact. We propose that the aromatic ring interaction can promote the peptide permeability by stabilizing a more compact structure of biphalin that would minimize the number of hydrogen bonds with water and therefore enhances partitioning into the model membrane.

Lipid membrane permeability of modified c[D-Pen2, D-Pen5]enkephalin peptides.

Permeability coefficients of a series of analogues of a potent opioid peptide, c[D-Pen2, D-Pen5]enkephalin, were measured in a model membrane system. The analogues included hydrophobic amino acid substitutions on position 3. Liposomes of a mixed composition consisting of zwitterionic lipids and cholesterol served as the model membranes. The obtained permeability coefficients range between 0.38 x 10(-12) and 2.9 x 10(-12) cm/s. These data were correlated with the hydrophobicity scale of Nozaki and Tanford (J. Biol. Chem. 246, 1971, 2211-2217) (correlation coefficient = 0.9933) and with determinations of lipid order perturbation by differential scanning calorimetry (correlation coefficient = -0.9779). The reasonably good correlation obtained within the family of analogues substituted on position 3 (Gly, Ala, Leu, Phe) indicates that changes in permeabilities are primarily related to increases in the partition coefficient of the peptide. However, Phe residue added on the N-terminal end of the peptide (position 0) does not appear to follow the observed trend, showing stronger lipid perturbation and lower permeability compared to the Phe3 analog. This observation demonstrates that each class of peptide modifications requires a new basis of permeability analysis and predictions.

Comparison of the membrane-bound states of two structurally similar delta-selective opioid peptides by transferred nuclear Overhauser effect spectroscopy and molecular modeling.

NMR spectroscopic, peptide-membrane conformational studies on [D-Pen2,D-Pen5]-enkephalin (DPDPE), an opioid receptor selective peptide, and an acyclic analog of DPDPE (DPDPE reduced at the disulfide bond) were conducted. The NMR method of transferred nuclear Overhauser effect (TRNOE) was used to obtain NOE profiles of the free and membrane bound forms of DPDPE and acyclic DPDPE. After comparison of the profiles of both peptides in the free and membrane-bound states, we hypothesize that the cyclic DPDPE undergoes little if any conformational change upon interaction with the membrane. However, for the acyclic analog, large changes in the NOE profile associated with backbone and side-chain groups were observed after interaction with the membrane. Results of computerized molecular modeling studies also were consistent with our theory that the free and membrane-bound forms of cyclic DPDPE have very similar free and membrane-bound states. The free acyclic DPDPE has a reverse turn conformation with sidechains situated so that hydrophobic surface exposure to aqueous solution is minimized. After membrane interaction, the acyclic DPDPE has an extended conformation near the carboxy terminus with aromatic sidechains widely separated. We propose that the interaction of the acyclic DPDPE with the membrane surface is mediated by the amino terminus. We further propose that the interaction of the cyclic DPDPE with the membrane surface is limited because the D-Pen2 side chain is covalently bonded and the aromatic side chains and backbone are only slightly altered after membrane contact. Permeability studies by Ramaswami et al. [(1992) Biochim. Biophys. Acta 1109(2), 195-202] demonstrated that the acyclic DPDPE permeated through membranes at a rate 4 times greater than cyclic DPDPE. We conclude that conformational and topographical flexibility may be critical factors in peptide-membrane interactions and permeability of bilayer membranes to opioid peptides.

Quasi-elastic light scattering determination of the size distribution of extruded vesicles.

The size distribution of phospholipid vesicles prepared by the freeze thaw-extrusion method were determined by the non-perturbing technique of quasi-elastic light scattering (QELS) and compared to latex particles of known size. Multiangle QELS experiments were performed to avoid errors due to the angular dependence of the scattering function of the particles. The experimentally determined autocorrelation function was analyzed by multiple mathematical procedures, i.e. single exponential, CUMULANT, exponential sampling, non-negatively constrained least square and CONTIN, in order to select suitable models for vesicle characterization. The most consistent results were obtained with CUMULANT, non-negatively constrained least square and CONTIN. In many instances single exponential analysis gave comparable results to these procedures, which indicates the vesicles have a narrow distribution of sizes. The influence of filter pore size, extrusion pressure and lipid concentration on the size and size distribution of extruded vesicles was determined. Extrusion through 100-, 200- and 400-nm pore size filters produced a unimodal distribution of vesicles, with somewhat smaller diameters as the extrusion pressure increased. The larger the filter pore size, the more dependent the vesicle size was on applied pressure. The observed vesicle size was independent of the lipid concentration between 0.1 and 10 mg ml-1.

Opioid peptide interactions with lipid bilayer membranes.

The interaction of the delta-opioid receptor selective peptides, cyclic [D-Pen2, D-Pen5]-enkephalin [DPDPE] and its acyclic analog, DPDPE(SH)2, with neutral phospholipid bilayer membranes was examined by permeability and calorimetry measurements. The permeabilities were accomplished by entrapping either peptide inside of unilamellar liposomes (composed of a mixture of a molar ratio 65:25:10 phosphatidylcholine/phosphatidylethanolamine/cholesterol) then monitoring the peptide efflux through the bilayer. The initial permeability of DPDPE (first 12 h) averaged over four experiments was (0.91 +/- 0.47).10(-12) cm s-1. In contrast the average permeability of the acylic DPDPE(SH)2 was (4.26 +/- 0.23).10(-12) cm s-1. The effect of these peptides on the phase transition, Tm, of 1,2-dipalmitoylphosphatidylcholine (DPPC) bilayers was examined by high sensitivity differential scanning calorimetry. The Tm, the calorimetric enthalpy, and the van 't Hoff enthalpy of DPPC were not significantly altered by the presence of DPDPE, whereas the calorimetric data for DPPC with DPDPE(SH)2 showed a small, yet significant, increase (0.2 degrees C) in the Tm with a 30% decrease in the cooperative unit. Both the permeability and calorimetry data reveal a stronger peptide-membrane interaction in the case of the more flexible acyclic peptide.

Photoinduced destabilization of liposomes.

The stability of two-component liposomes composed of the polymerizable 1,2-bis-[10-(2',4'-hexadienoyloxy)decanoyl]-sn-glycero-3-phosphati dylcholine (SorbPC) and either a phosphatidylethanolamine (PE) or a phosphatidylcholine (PC) were examined via fluorescence leakage assays. Ultraviolet light exposure of SorbPC-containing liposomes forms poly-SorbPC, which phase separates from the remaining monomeric lipids. If the nonpolymerizable lipids are PE's, then the photoinduced polymerization destabilizes the liposome with loss of aqueous contents. The permeability of the control dioleoylPC/SorbPC membranes was not affected by photopolymerization of SorbPC. The photodestabilization of dioleoylPE/SorbPC (3:1) liposomes required the presence of oligolamellar liposomes. NMR spectroscopy of extended bilayers of dioleoylPE/SorbPC (3:1) showed that the photopolymerization lowers the temperature for the appearance of 31P NMR signals due to the formation of isotropically symmetric lipid structures. These observations suggest the following model for the photoinduced destabilization of liposomes composed of PE/SorbPC; photopolymerization induced phase separation with the formation of enriched domains of PE, which allows the close approach of apposed regions of enriched PE lamellae and permits the formation of an isotropically symmetric structure between the lamellae. The formation of such an interlamellar attachment (ILA) between the lamellae of an oligolamellar liposome provides a permeability pathway for the light-stimulated leakage of entrapped water-soluble reagents.