Lateral diffusion of gramicidin S, M-13 coat protein and glycophorin in bilayers of saturated phospholipids. Mean field and Monte Carlo studies

We have developed a general model that relates the lateral diffusion coefficient of one isolated large intrinsic molecule (mol. wt. greater than or approximately 1000) in a phosphatidylcholine bilayer to the static lipid hydrocarbon chain order. We have studied how protein lateral diffusion can depend upon protein-lipid interactions but have not investigated possible non-specific contributions from gel-state lattice defects. The model has been used in Monte Carlo simulations or in mean-field approximations to study the lateral diffusion coefficients of Gramicidin S, the M-13 coat protein and glycophorin in dimyristoyl- and dipalmitoylphosphatidylcholine (DMPC and DPPC) bilayers as functions of temperature. Our calculated lateral diffusion coefficients for Gramicidin S and the M-13 coat protein are in good agreement with what has been observed and suggest that Gramicidin S is in a dimeric form in DMPC bilayers. In the case of glycophorin we find that the 'ice breaker' effect can be understood as a consequence of perturbation of the lipid polar region around the protein. In order to understand this effect is necessary that the protein hydrophilic section perturb the polar regions of at least approx. 24 lipid molecules, in good agreement with the numbers of 29-30 measured using 31P-NMR. Because of lipid-lipid interactions this effect extends itself out to four or five lipid layers away from the protein so that the hydrocarbon chains of between approx. 74 and approx. 108 lipid molecules are more disordered in the gel phase, so contributing less to the transition enthalpy, in agreement with the numbers of 80-100 deduced from differential scanning calorimetry (DSC). An understanding of the abrupt change in the diffusion coefficient at a temperature below the main bilayer transition temperature requires an additional mechanism. We propose that this change may be a consequence of a 'coupling-uncoupling' transition involving the protein hydrophilic section and the lipid polar regions, which may be triggered by the lipid bilayer pretransition. Our calculation of the average number of gauche bonds per lipid chain as a function of temperature and distance away from an isolated polypeptide or integral protein shows the extent of statically disordered lipid around such molecules. The range of this disorder depends upon temperature, particularly near the main transition.

Thallium-205 nuclear magnetic resonance study of the thallium(I)-gramicidin A association in trifluoroethanol

Association of the thallous ion with gramicidin in 2,2,2-trifluoroethanol has been investigated by thallium-205 NMR spectroscopy. The data obtained suggest that the gramicidin dimer has two strong binding sites and one or more weak binding sites. Association constants for the strong binding sites were found to have the same value. From the temperature dependence of the strong binding site association constants, values for the association enthalpy and entropy of -2.13 +/- 0.12 kcal/mol and +5.45 +/- 0.04 eu, respectively, were obtained.

High-resolution NMR studies of transmembrane cation transport: use of an aqueous shift reagent for 23Na

23Na NMR studies of large unilamellar vesicles of egg lecithin in salt solutions are reported. A shift reagent, the dysprosium nitrilotriacetate ion Dy[N(CH2CO2)3]3-2 has been used to distinguish between 23Na+ inside and outside the vesicles. When both are present and the shift reagent is present on only one side, two clearly distinct resonances are observed. Creation of a Na+ concentration gradient and subsequent catalysis of passive transport induced by the introduction of gramicidin can be monitored easily by using the relative intensities of the two resonances. We report the observation of transport both out of and into vesicles.

Synthesis and characterization of 1-(13) C-D X Leu12, 14 gramicidin A

The 13C-D-Leu12, 14 gramicidin A was synthesized by the solid phase method incorporating 13C-D-leucine in positions 12 and 14 with about 25 and 50% enrichment, respectively. The pentadecapeptide was removed from the resin by ethanolamine treatment, with the N-protecting group (Boc) still on. After removal of the protecting group, the peptide was formylated and purified by preparative t.l.c. to obtain 13C-D-Leu12, 14 gramicidin A in a very pure state in an overall yield of about 12.5%. The peptide was then thoroughly characterized by HPLC which gave one single peak with the same retention time as that of Val1-gramicidin A of the natural gramicidin mixture. The CD spectra of the synthetic and the HPLC purified natural Val1-GA were obtained and found to be identical, indicating the optical purity of the sample. The synthetic GA was characterized by 13C n.m.r. spectrum and compared with that of natural GA. Single channel conductance parameters of the synthetic GA were determined and found to be indistinguishable from those of natural Val1-GA in lipid bilayer membranes and the mean channel lifetime was found to be as reported earlier by others.

Location of monovalent cation binding sites in the gramicidin channel

Six syntheses of gramicidin A have been carried out, each with 90% 13 C enrichment of a single carbonyl carbon these being the formyl, Val-1, Trp-9, Trp-11, Trp-13, and Trp-15 carbonyl carbons. Each gramicidin A was incorporated as the channel state into phospholipid structures, and the chemical shift of the carbonyl carbon resonance was monitored by 13C NMR as a function of ion concentration. Plots of Na+- and Tl+-induced chemical shifts as a function of carbonyl location in the channel indicate two symmetrically related binding sites centered at the tryptophan carbonyls and separated by 23 A. The absence of ion-induced chemical shifts for the formyl and Val-1 carbonyl carbon resonances indicates that there is no binding site midway through the channel but rather a central free-energy barrier for ion transit through the channel. Ion induced chemical shifts of the tryptophan carbonyl carbon resonances at 100 mM Na+ verify that the tight binding constant (Kbt congruent to 70 M-1), observed with 23Na NMR, results from binding within the channel. This observation and the lateral, triangular distribution of the coordinating Trp-9, -11, and -13 carbonyls combine to provide an experimental demonstration that the carbonyls of the walls of the channel directly coordinate the ion, successfully competing with the polar solvent. With the binding sites verified and localized, it is possible to conclude that the transport mechanism for Na+ is well represented by the case of the two-site model [D. W. Urry, Venkatachalam, C. M., Spisni, A., Läuger, P. & Khaled, M. A. (1980) Proc. Natl. Acad. Sci. USA 77, 2028--2032].

Intrinsic proteins and their effect upon lipid hydrocarbon chain order

We present evidence that at temperatures greater than their main transition temperature, phospholipid molecules that are trapped within clusters of intrinsic molecules such as polypeptides or proteins have the ends of their hydrocarbon chains more statically disordered than those of lipid molecules far from such intrinsic molecules. We have constructed a model in which the lipids are divided into three populations: (i) those that are not adjacent to any protein ("free" lipids), (ii) those that are adjacent to only one protein ("adjacent" lipids), and (iii) those that are "trapped" between two or three proteins. We applied this model to study deuterium nuclear magnetic resonance of dimyristoyl-3-sn-phosphatidylcholine (DMPC) bilayers containing gramicidin A' or cytochrome oxidase and found that while the methyl groups of adjacent lipids are slightly more statically ordered than those of free lipids, the methyl groups of trapped lipids are more statically disordered than those of free lipids. We propose a physical explanation for this and show that phosphorus-31 nuclear magnetic resonance data for DMPC-cytochrome oxidase bilayers can be understood as a consequence of changes in the polar region of only trapped lipids.

Single channel properties of D-Leu2-gramicidin A. Side chain modulation of channel lifetime

The channel forming properties of synthetic gramicidin A and DLeu2-gramicidin A were compared in black lipid membranes. The most probable single channel conductance was identical for both derivatives but in each case a distribution of smaller channel sizes was observed. However, the lifetime of the channel formed by DLeu2-gramicidin A was considerably shorter than for gramicidin A. The DLeu2 substitution is considered to interfere with the head to head hydrogen bonding which forms the conducting dimer, thus destabilizing the dimeric structure of the channel and reducing the lifetime. This represents the first demonstration of side-chain modulation of channel lifetime.

Solution dimensions of the gramicidin dimer by dynamic light scattering

Gramicidin is thought to form dimeric helical rods in alcohol solutions. In addition, there is evidence that the rod dimensions change upon addition of potassium ions. The present work reports values for the translational and rotational diffusion coefficients of gramicidin in methanol and 95% ethanol and in these same solvents with added KSCN. Solution dimensions are calculated from the diffusion coefficients. The results suggest that gramicidin exists primarily as dimers in these solutions and that the gramicidin rod does indeed become shorter upon addition of potassium ion. These results are consistent with those obtained from X-ray studies on single crystals grown from alcohol solutions.

Binding constants of Li+, K+, and Tl+ in the gramicidin channel determined from water permeability measurements

In an open circuit there can be no net cation flux through membranes containing only cation-selective channels, because electroneutrality must be maintained. If the channels are so narrow that water and cations cannot pass by each other, then the net water flux through those "single-file" channels that contain a cation is zero. It is therefore possible to determine the cation binding constants from the decrease in the average water permeability per channel as the cation concentration in the solution is increased. Three different methods were used to determine the osmotic water permeability of gramicidin channels in lipid bilayer membranes. The osmotic water permeability coefficient per gramicidin channel in the absence of cations was found to be 6 x 10(-14) cm3/s. As the cation concentration was raised, the water permeability decreased and a binding constant was determined from a quantitative fit to the data. When the data were fitted assuming a maximum of one ion per channel, the dissociation constant was 115 mM for Li+, 69 mM for K+, and 2 mM for Tl+.

Water transport and ion-water interaction in the gramicidin channel

The diffuse permeability and the diffusion coefficient of water (Dw) in the gramicidin channel is determined from the osmotic water permeability of the channel and "single file" pore theory. Dw is about 7% of the self-diffusion coefficient of bulk water. The diffusion coefficient of a single water molecule alone in the channel is also determined and is about equal to the value in bulk water. This provides an estimate of the mobility of water on the channel walls in the absence of water-water interaction. Since the gramicidin channel walls should be representative of uncharged polar protein surfaces, this result provides direct evidence that the presence of a cation in the channel reduces the hydraulic water permeability by a factor ranging from 60 for Tl+ to 5 for Na+. The diffusion coefficient of a cation (Dc) in the channel is estimated and compared with Dw. For Na+ it is found that Dc approximately equal to Dw, which implies that the movement of the row of water molecules through the channel determines the local mobility of Na+. Thus, it seems that short range ion-wall interactions are not important in determining the channel conductance for Na+. In contrast, for Li+, local ion-wall interactions probably do limit the conductance.

Energy barriers for passage of ions through channels. Exact solution of two electrostatic problems

Exact solutions are given to two electrostatic problems relevant to ion permeation through pores in membranes. The first assesses the importance of the pore forming molecule as a dielectric shield. It is shown on the basis of structural and dielectric considerations alone (neglecting effects attributable to possible charge distribution at the interior surface of the pre-former) that the minimum electrostatic barrier for monovalent ion passage through a gramicidin-like channel is 11 kT. It is further shown that given favorable circumstances, dielectric shielding might dramatically reduce the barrier to ion passage through potassium channels. The second problem considers the error introduced by treating ions as point charges. It is shown that for structureless pores the point charge approximation introduces no meaningful error, even if the ratio of ion radius to pore radius is as great as 0.95.

The gramicidin A channel: a review of its permeability characteristics with special reference to the single-file aspect of transport

Gramicidin A forms univalent cation-selective channels of approximately 4 A diameter in phospholipid bilayer membranes. The transport of ions and water throughout most of the channel length is by a single-file process; that is, cations and water molecules cannot pass each other within the channel. The implications of this single-file mode of transport for ion movement are considered. In particular, we show that there is no significant electrostatic barrier to ion movement between the energy wells at the two ends of the channel. The rate of ion translocation (e.g., Na+ or Cs+) through the channel between these wells is limited by the necessity for an ion to move six water molecules in single file along with it; this also limits the maximum possible value for channel conductance. At all attainable concentrations of NaCl, the gramicidin A channel never contains more than one sodium ion, whereas even at 0.1 M CsCl, some channels contain two cesium ions. There is no necessity to postulate more than two ion-binding sites in the channel or occupancy of the channel by more than two ions at any time.

The dependence of the conductance and lifetime of gramicidin channels on the thickness and tension of lipid bilayers

The lifetimes of channels formed by natural gramicidin and its dimeric analog in monoglyceride lipid bilayers of various compositions were investigated. The bilayer surface tension was altered by changing the length of the monoglycerides' fatty acid chain or the chain length of hydrocarbon solvent by isomerization or saturation of the lipid, by varying the amount of solvent in the bilayer, and by changing the salt composition of the aqueous solutions. The logarithms of mean channel lifetimes were found to be proportional to the surface tension of the membrane irrespective of how the surface tension was changed. In contrast, no simple relationship between channel conductance and surface tension or bilayer thickness was found.

[Study of the volt-ampere characteristics of ion channels by transmembrane current harmonics]

A method is proposed for measuring the coefficient of non--linearity beta of current--voltage characteristics of the class i (U) approximately U (1 + beta U2) of ionic channels formed by grammicidine A (Gra) and amphotericine B by the 3rd harmonics of the membrane current. For Gra A beta depends on the concentration of electrolyte c increasing lg c from -17 B-2 at 0.03 M to 8 B-2 at 3.4 M KCl turning to 0 at c0 = 0.3 divided by 1 M. The membranes of egg lecithin and glycerylmonooleate (GMO) differ in c0 value. Substitution of K+ ion for Li+, of the membrane solvent (n-heptane for n-hexadecane) and freezing of the GMO membrane do not affect beta.

Intrinsic molecules in fluid phospholipid bilayers. Fluorescence probe studies

Fluorescence probe data using 1,6-diphenyl-1,3,5-hexatriene for various concentrations of intrinsic molecules (cholesterol, gramicidin A amd cytochrome oxidase) within fluid lipid bilayers have been examined. The polarization value increases with increasing concentration of intrinsic molecule and then approaches a limiting value. Empirical curve-fitting of the experimental data, change of polarization with concentration, shows that each system can be fitted approximately by an exponential curve. A theory has been constructed based upon the assumption that only one intrinsic molecule need be adjacent to a fluorescent probe molecule to affect its motion drastically. The change in probe motion then depends upon the probability p of all positions next to a lipid chain being free of intrinsic molecules. The value of the probability p has been calculated and it is shown that (formula: see text) depending on whether the intrinsic molecule spans the lipid bilayer or not. The approximation p = e-Mx gives a good fit to the data for all x, thereby explaining the observed phenomenological fit. The fluorescent probe data is interpreted to show that protein-protein contacts increase as the intrinsic protein concentration increases within the lipid bilayer. An apparent dichotomy between the results from the fluorescence probe and from the deuterium magnetic resonance is explained in terms of a dominant affect on the probe being its hindrance to motion by interaction with the intrinsic molecule (protein) whilst individual C2H2 groups of the chain may exhibit greater disorder.

Fusion of small unilamellar liposomes with phospholipid planar bilayer membranes and large single-bilayer vesicles

Small unilamellar phosphatidylserine/phosphatidylcholine liposomes incubated on one side of planar phosphatidylserine bilayer membranes induced fluctuations and a sharp increase in the membrane conductance when the Ca2+ concentration was increased to a threshold of 3--5 mM in 100 mM NaCl, pH 7.4. Under the same ionic conditions, these liposomes fused with large (0.2 micrometer diameter) single-bilayer phosphatidylserine vesicles, as shown by a fluorescence assay for the mixing of internal aqueous contents of the two vesicle populations. The conductance behavior of the planar membranes was interpreted to be a consequence of the structural rearrangement of phospholipids during individual fusion events and the incorporation of domains of phosphatidylcholine into the Ca2+-complexed phosphatidylserine membrane. The small vesicles did not aggregate or fuse with one another at these Ca2+ concentrations, but fused preferentially with the phosphatidylserine membrane, analogous to simple exocytosis in biological membranes. Phosphatidylserine vesicles containing gramicidin A as a probe interacted with the planar membranes upon raising the Ca2+ concentration from 0.9 to 1.2 mM, as detected by an abrupt increase in the membrane conductance. In parallel experiments, these vesicles were shown to fuse with the large phosphatidylserine liposomes at the same Ca2+ concentration.

Use of nuclear Overhauser effect in the study of peptides and proteins

The nuclear Overhauser effect (NOE) leads to changes in the intensity of signal(s) of a set of nuclei as a function of their respective distances. The use of NOE allows to obtain structural informations on peptides and proteins in solution as well as the study of interactions between small ligands and biomolecules. In this review, aspects of the basic theory of the NOE will be presented and the more recent applications of homonuclear and heteronuclear NOE's in biomolecules will be surveyed. Typical examples will be illustrated and limitations of the method will be discussed.

Ion movements in gramicidin pores. An example of single-file transport

Experimental results on ion movement through gramicidin membrane channels are presented and discussed in terms of ion transport in the simplest single-file pore (for review see Urban, B.W. and Hladky, S.B. (1979) Biochim. Biophys. Acta 554, 410-429). Single-channel conductance and bi-ionic potential data for Na+, K+, Cs+, NH4+ and Tl+ are used to assign values to the rate constants of the model. Not all of the rate constants can be determined uniquely and simplifications are introduced to reduce the number of free parameters. The simplified model gives good quantitative fits to the experimental results for Na+, K+, Cs+ and NH4+. For Tl+, although the model accounts qualitatively for the salient features of the results, the quantitative agreement is less satisfactory. Predictions calculated from the model and the fitted rate constants are compared with independent data from blocking and tracer-flux measurements. In agreement with experiment, the model shows that only Tl+ blocks the Na+ conductance significantly. Furthermore, the exponent, n, in the tracer flux ratio rises, as observed, well above unity. The values for the rate constants suggest internal consistency of the model in that entry is always slower to singly occupied pores than to empty pores while exit is always faster from doubly as compared to singly occupied pores. The agreement between model prediction and experimental results suggests that the main features of ion transport in the gramicidin channel arise from cation-cation interaction in a single-file pore.