Structural features that modulate the transmembrane migration of a hydrophobic peptide in lipid vesicles.

Two approaches employing nuclear magnetic resonance (NMR) were used to investigate the transmembrane migration rate of the C-terminal end of native alamethicin and a more hydrophobic analog called L1. Native alamethicin exhibits a very slow transmembrane migration rate when bound to phosphatidylcholine vesicles, which is no greater than 1 x 10(-4) min(-1). This rate is much slower than expected, based on the hydrophobic partition energies of the amino acid side chains and the backbone of the exposed C-terminal end of alamethicin. The alamethicin analog L1 exhibits crossing rates that are at least 1000 times faster than that of native alamethicin. A comparison of the equilibrium positions of these two peptides shows that L1 sits approximately 3-4 A deeper in the membrane than does native alamethicin (Barranger-Mathys and Cafiso. 1996. Biochemistry. 35:489). The slow rate of alamethicin crossing can be explained if the peptide helix is irregular at its C-terminus and hydrogen bonded to solvent or lipid. We postulate that L1 does not experience as large a barrier to transport because its C-terminus is already buried within the membrane interface. This difference is most easily explained by conformational differences between L1 and alamethicin rather than differences in hydrophobicity. The results obtained here demonstrate that side-chain hydrophobicity alone cannot account for the energy barriers to peptide and protein transport across membranes.

1H, 15N and 13C resonance assignments and secondary structure determination of the RNA-binding domain of E.coli rho protein.

Protein fragments containing the RNA-binding domain of Escherichia coli rho protein have been over-expressed in E. coli. NMR spectra of the fragment containing residues 1-116 of rho protein (Rho116) show that a region of this protein is unfolded in solution. Addition of (dC)10 to this fragment stabilizes the folded form of the protein. The fragment comprising residues 1-130 of rho protein (Rho130) is found to be stably folded, both in absence and presence of nucleic acid. NMR studies of the complex of Rho130 with RNA and DNA oligonucleotides indicate that the binding-site size, affinity, and specificity of Rho130 are similar to those of intact rho protein; therefore, Rho130 is a suitable model of the RNA-binding domain of Rho protein. NMR line widths as well as titration experiments of Rho130 complexed with oligonucleotides of various lengths suggests that Rho130 forms oligomers in the presence of longer oligonucleotides. 1H, 15N and 13C resonance assignments were facilitated by the utilization of two pulse sequences, CN-NOESY and CCH-TOCSY. The secondary structure of unliganded Rho130 has been determined by NMR techniques, and it is clear that the RNA-binding domain of rho is more structurally similar to the cold shock domain than to the RNA recognition motif.

Distribution of general anesthetics in phospholipid bilayers determined using 2H NMR and 1H-1H NOE spectroscopy.

The effect of the general anesthetics halothane, enflurane, and isoflurane on hydrocarbon chain packing in palmitoyl(d31)oleoylphosphatidylcholine membranes in the liquid crystalline phase was investigated using 2H NMR. Upon the addition of the anesthetics, the first five methylene units near the interface generally show a very small increase in segmental order, while segments deeper within the bilayer show a small decrease in segmental order. From the 2H NMR results, the chain length for the perdeuterated palmitoyl chain in the absence of anesthetic was found to be 12.35 A. Upon the addition of halothane, enflurane, or isoflurane, the acyl chain undergoes slight contractions of 0.11, 0.20, or 0.16 A, respectively, at 50 mol % anesthetic. A simple model was used to estimate the relative amounts of anesthetic located near the interface and deeper in the bilayer hydrocarbon region, and only a slight preference for an interfacial location was observed. Intermolecular 1H-1H nuclear Overhauser effects (NOEs) were measured between phospholipid and halothane protons. These NOEs are consistent with the intramembrane location of the anesthetics suggested by the 2H NMR data. In addition, the NOE data indicate that anesthetics prefer the interfacial and hydrocarbon regions of the membrane and are not found in high concentrations in the phospholipid headgroup.

Structure of micelle-associated alamethicin from 1H NMR. Evidence for conformational heterogeneity in a voltage-gated peptide.

Alamethicin is a 20 amino acid peptide that produces a voltage-dependent conductance in membranes. To understand the mechanism by which this peptide becomes voltage-gated, the structure of alamethicin bound to micelles was examined using high-resolution 1H nuclear magnetic resonance (NMR). Two-dimensional correlation and nuclear Overhauser effect spectroscopy (NOESY) were carried out on alamethicin incorporated into perdeuterated sodium dodecyl sulfate (SDS) micelles, and the 1H NMR spectrum of the peptide in micelles was assigned. The intensities of the HN-HN(i,i+1), H alpha-HN(i,i+1), H alpha-NH(i,i+3), H alpha-H beta (i,i+3), and H alpha-NH(i,i+4) cross peaks in the NOESY spectrum suggest that the N-terminal half of the peptide is predominantly alpha-helical, while the C-terminal half has a less regular or more flexible structure. The exposure of micelle bound alamethicin to the aqueous solution was determined by examining the effect of aqueous paramagnetic reagents on the line widths of the peptide protons. These measurements suggest that alamethicin is buried in the micelle. A set of restraints consisting of 175 distances (derived from NOESY spectra), five dihedral angles, and two hydrogen bond distances were used in a simulated annealing procedure that yielded structures for micelle associated alamethicin. The structures that were generated with simulated annealing were largely helical from residues 4-9 and 12-16. A limited number of structural forms were obtained. The main difference among forms involved the backbone conformations of MeA10, Gly11, and Leu12 and resulted in structures that were straight or had different amounts of bend. The structural forms could be easily interconverted by rotation of the psi and phi angles of residues 10-12. The rotational freedom at or near MeA10 may be a result of Pro14, which would be the normal hydrogen-bonding position for the peptide carbonyl of MeA10. These results suggest that conformation rearrangements at or near MeA10 may play a role in the voltage-gating of alamethicin.

Determination of the molecular dynamics of alamethicin using 13C NMR: implications for the mechanism of gating of a voltage-dependent channel.

Alamethicin is a channel-forming peptide antibiotic that produces a highly voltage-dependent conductance in planar bilayers. To provide insight into the mechanisms for its voltage dependence, the dynamics of the peptide were examined in solution using nuclear magnetic resonance. Natural-abundance 13C spin-lattice relaxation rates and 13C-1H nuclear Overhauser effects of alamethicin were measured at two magnetic field strengths in methanol. This information was interpreted using a model-free approach to obtain values for the overall correlation times as well as the rates and amplitudes of the internal motions of the peptide. The picosecond, internal motions of alamethicin are highly restricted along the peptide backbone and indicate that it behaves as a rigid helical rod in solution. The side chain carbons exhibit increased segmental motion as their distance from the peptide backbone is increased; however, these motions are not unrestricted. Methyl group dynamics are also consistent with the restricted motions observed for the backbone carbons. There is no evidence from these dynamics measurements for a hinged motion of the peptide about proline-14. Alamethicin appears to be slightly less structured in methanol than in the membrane; as a result, alamethicin is also expected to behave as a rigid helix in the membrane. This suggests that the gating of this peptide involves changes in the orientation of the entire helix, rather than the movement of a segment of the peptide backbone.

Comparison of the lipid acyl chain dynamics between small and large unilamellar vesicles.

13C NMR spin-lattice relaxation (T1) rates and 13C-1H nuclear Overhauser effects (NOEs) were measured in an identical fashion in two lipid preparations having dramatically different curvatures. The T1 times that were obtained at four magnetic field strengths were fit along with the NOEs to simple models for lipid molecular dynamics. The results indicate that phospholipid chain ordering and dynamics are virtually identical in small and large unilamellar vesicles at the time scales sampled by these 13C-NMR studies. The order parameters and reorientational correlation times that characterize the amplitudes and rates of internal acyl chain motions were equal within experimental error for the methylene segments in the middle of the chains. The only significant differences in order parameters and correlation times between the two vesicle types were small and appeared at the ends of the acyl chains. At the carbonyl end the order was slightly higher in small vesicles than large vesicles, and at the methyl end the order was slightly lower for small vesicles. This indicates that in the more planar systems the acyl chains exhibit a slightly flatter order profile than in more highly curved membranes. The use of the same experimental approach in both small and large vesicle systems provided a more reliable and accurate assessment of the effect of curvature on molecular order than has been previously obtained.

Dynamics and aggregation of the peptide ion channel alamethicin. Measurements using spin-labeled peptides.

Two spin-labeled derivatives of the ion conductive peptide alamethicin were synthesized and used to examine its binding and state of aggregation. One derivative was spin labeled at the C-terminus and the other, a leucine analogue, was spin labeled at the N-terminus. In methanol, both the C and N terminal labeled peptides were monomeric. In aqueous solution, the C-terminal derivative was monomeric at low concentrations, but aggregated at higher concentrations with a critical concentration of 23 microM. In the membrane, the C-terminal label was localized to the membrane-aqueous interface using 13C-NMR, and could assume more than one orientation. The membrane binding of the C-terminal derivative was examined using EPR, and it exhibited a cooperativity seen previously for native alamethicin. However, this cooperativity was not the result of an aggregation of the peptide in the membrane. When the spectra of either the C or N-terminal labeled peptide were examined over a wide range of membrane lipid to peptide ratios, no evidence for aggregation could be found and the peptides remained monomeric under all conditions examined. Because electrical measurements on this peptide provide strong evidence for an ion-conductive aggregate, the ion-conductive form of alamethicin likely represents a minor fraction of the total membrane bound peptide.

Localizing the nitroxide group of fatty acid and voltage-sensitive spin-labels in phospholipid bilayers.

The intramembrane locations of several spin labeled probes in small egg phosphatidylcholine (egg PC) vesicles were determined from the enhancement of the 13C nuclear spin lattice relaxation of the membrane phospholipid. Electron paramagnetic resonance (EPR) spectroscopy was also used to measure the relative environmental polarities of the spin labels in egg PC vesicles, ethanol and aqueous solution. The binding location of the spin label group was determined for a pair of hydrophobic ion spin labels, a pair of long chain amphiphiles, and three stearates containing doxyl groups at the 5, 10 and 16 positions. The nuclear relaxation results indicate that the spin label groups on the stearates are located nearer to the membrane exterior than the analogous positions of the unlabeled phospholipid acyl chains. In addition, the spin label groups of the hydrophobic ions and long chain amphiphiles are located near the acyl chain methylene immediately adjacent to the carboxyl group. The relative polarities, determined by the EPR technique, are consistent with the nuclear relaxation results. This information, when combined with information on their electrical properties, allows for an assessment of the conformation and position of these voltage sensitive probes in membranes.

Localization of hydrophobic ions in phospholipid bilayers using 1H nuclear Overhauser effect spectroscopy.

The binding location for the hydrophobic ions tetraphenylphosphonium (TPP+) and tetraphenylboron (TPB-) was studied in sonicated phosphatidylcholine (PC) vesicles by measuring time-dependent and steady-state intermolecular 1H nuclear Overhauser effects (NOE's). Intermolecular cross-relaxation was also investigated by two-dimensional NOE spectroscopy. Information on the distance and order parameter dependence of the NOE's was obtained from a simple simulation of the NOE's in the alkyl chain region. Taken together, the NOE data and the simulation provide strong evidence that TPB- and TPP+, at low concentrations (less than or equal to 10 mol%), are localized in the alkyl chain region of the bilayer. At these lower concentrations of TPP+ or TPB-, no significant effect on lipid 13C T1 or T2 relaxation rates is detected. The proposed location is consistent with the expected free energy profiles for hydrophobic ions and with the carbonyl oxygens or interfacial water as the source of the membrane dipole potential. At higher ion/lipid ratios (greater than or equal to 20 mol%), TPB-/lipid NOE's increase. This results from a specific association of TPB- with the choline head group.

31P NMR spectra of rod outer segment and sarcoplasmic reticulum membranes show no evidence of immobilized components due to lipid-protein interactions.

31P NMR studies of rod outer segment (ROS) and sarcoplasmic reticulum (SR) membranes have been performed under conditions where broad and narrow spectral components can be clearly resolved. Control studies of an anhydrous, solid powder of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), as well as aqueous binary mixtures of 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) and 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), demonstrate clearly that broad spectral components can be detected. For the codispersions of DSPC and DOPC in the mixed-phase region at 22 degrees C, the 31P NMR spectra consist of a superposition of a broad component and a narrow, axially symmetric component, due to coexisting solid and liquid-crystalline domains, which are in slow exchange on the 31P NMR time scale. The 31P NMR spectra of the native ROS and SR membranes, however, consist of only a narrow component, to within experimental error, indicating that most or all of the phospholipids are in the liquid-crystalline (L alpha) phase at 22 degrees C. The above conclusions are in agreement with many, but not all, previous studies [see, e.g., Yeagle, P.L. (1982) Biophys. J. 37, 227-239]. It is estimated that at most 10% of the phospholipids in the ROS and SR membranes could give rise to broad 31P NMR spectral components, similar to those seen for anhydrous or solid-phase lipids, corresponding to approximately 7 phospholipids/rhodopsin molecule and approximately 11 phospholipids/Ca2+-ATPase molecule, respectively.

Lipid-protein interactions in reconstituted membranes containing acetylcholine receptor.

Functional membranes containing purified Torpedo californica acetylcholine receptor and dioleoylphosphatidylcholine (DOPC) were prepared by a cholate dialysis procedure with lipid to protein ratios of 100-400 to 1 (mol/mol). Spin-labeled lipids were incorporated into the reconstituted membranes and into native membranes prepared from Torpedo electroplax, and electron paramagnetic resonance (EPR) spectra were recorded between 0 and 20 degrees C. The spin-labels included nitroxide derivatives of stearic acid (16-doxylstearic acid), androstane, phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylserine (PS), and phosphatidic acid (PA). The phospholipid spin-labels had 16-doxylstearic acid in the sn-2 position. All the spectra showed two components corresponding to a relatively mobile bilayer component and a motionally restricted "protein-perturbed" component. The relative amounts of mobile and perturbed components were quantitated by spectral subtraction and integration techniques. The mobile/perturbed ratio was somewhat temperature dependent, and the results are discussed in terms of exchange between mobile and perturbed environments. Plots of the mobile/perturbed ratios vs. lipid/protein ratios at 1 degree C gave straight lines from which the relative binding affinity of each spin-label and the number of perturbed lipids per receptor protein could be calculated. All the spin-labels gave similar values for the number of perturbed lipids (40 +/- 7), a number close to the number of lipids that will fit around the intramembranous perimeter of the receptor. The affinities of the spin-labeled lipids for the receptor relative to DOPC were androstane (K = 4.3) congruent to 16-doxylstearic acid (4.1) greater than PA (2.7) greater than PE (1.1) approximately PC (1.0) approximately PS (0.7). The lipids having the highest affinity for the acetylcholine receptor were also those that have the largest effects on the ion flux functional properties of the receptor, and the results are discussed in terms of lipid effects on receptor function.

Lipid bilayer dynamics and rhodopsin-lipid interactions: new approach using high-resolution solid-state 13C NMR.

High-resolution, solid-state 13C NMR spectra have been obtained for unsonicated multilamellar dispersions of 1,2-dilauryl-sn-glycero-3-phosphocholine (DLPC), recombinant membranes containing DLPC and rhodopsin, and native retinal rod disk membranes. The roles of 1H dipolar decoupling, 1H-13C cross-polarization, and magic-angle sample spinning have been investigated. Rotating-frame 13C relaxation times have been measured and are discussed in terms of lipid bilayer dynamics and rhodopsin-lipid interactions.