Dynamical and temperature-dependent effects of lipid-protein interactions. Application of deuterium nuclear magnetic resonance and electron paramagnetic resonance spectroscopy to the same reconstitutions of cytochrome c oxidase

Interaction of detergents with biological membranes: Comparison of fluorescence assays with filtration protocols and implications for the rates of detergent association, dissociation and flip-flop

The present study mainly consists of a re-evaluation of the rate at which C₁₂E₈, a typical non-ionic detergent used for membrane studies, is able to dissociate from biological membranes, with sarcoplasmic reticulum membrane vesicles being used as an example. Utilizing a brominated derivative of C₁₂E₈ and now stopped-flow fluorescence instead of rapid filtration, we found that the rate of dissociation of this detergent from these membranes, merely perturbed with non-solubilizing concentrations of detergent, was significantly faster (t_1/2 < 10 ms) than what had previously been determined (t_1/2 ~300–400 ms) from experiments based on a rapid filtration protocol using ¹⁴C-labeled C₁₂E₈ and glass fiber filters (Binding of a non-ionic detergent to membranes: flip-flop rate and location on the bilayer, by Marc le Maire, Jesper Møller and Philippe Champeil, Biochemistry (1987) Vol 26, pages 4803–4810). We here pinpoint a methodological problem of the earlier rapid filtration experiments, and we suggest that the true overall dissociation rate of C₁₂E₈ is indeed much faster than previously thought. We also exemplify the case of brominated dodecyl-maltoside, whose kinetics for overall binding to and dissociation from membranes comprise both a rapid and a sower phase, the latter being presumably due to flip-flop between the two leaflets of the membrane. Consequently, equilibrium is reached only after a few seconds for DDM. This work thereby emphasizes the interest of using the fluorescence quenching associated with brominated detergents for studying the kinetics of detergent/membrane interactions, namely association, dissociation and flip-flop rates.

Membrane solid-state NMR in Canada: A historical perspective

This manuscript presents an overview of more than 40years of membrane solid-state nuclear magnetic resonance (NMR) research in Canada. This technique is a method of choice for the study of the structure and dynamics of lipid bilayers; bilayer interactions with a variety of molecules such as membrane peptides, membrane proteins and drugs; and to investigate membrane peptide and protein structure, dynamics, and topology. Canada has a long tradition in this field of research, starting with pioneering work on natural and model membranes in the 1970s in a context of emergence of biophysics in the country. The 1980s and 1990s saw an emphasis on studying lipid structures and dynamics, and peptide-lipid and protein-lipid interactions. The study of bicelles began in the 1990s, and in the 2000s there was a rise in the study of membrane protein structures. Novel perspectives include using dynamic nuclear polarization (DNP) for membrane studies and using NMR in live cells. This article is part of a Special Issue entitled: Biophysics in Canada, edited by Lewis Kay, John Baenziger, Albert Berghuis and Peter Tieleman.

Self-Assembly of X-Shaped Bolapolyphiles in Lipid Membranes: Solid-State NMR Investigations

A novel class of rigid-rod bolapolyphilic molecules with three philicities (rigid aromatic core, mobile aliphatic side chains, polar end groups) has recently been demonstrated to incorporate into and span lipid membranes, and to exhibit a rich variety of self-organization modes, including macroscopically ordered snowflake structures with 6-fold symmetry. In order to support a structural model and to better understand the self-organization on a molecular scale, we here report on proton and carbon-13 high-resolution magic-angle spinning solid-state NMR investigations of two different bolapolyphiles (BPs) in model membranes of two different phospholipids (DPPC, DOPC). We elucidate the changes in molecular dynamics associated with three new phase transitions detected by calorimetry in composite membranes of different composition, namely, a change in π-π-packing, the melting of lipid tails associated with the superstructure, and the dissolution and onset of free rotation of the BPs. We derive dynamic order parameters associated with different H-H and C-H bond directions of the BPs, demonstrating that the aromatic cores are well packed below the final phase transition, showing only 180° flips of the phenyl ring, and that they perform free rotations with additional oscillations of the long axis when dissolved in the fluid membrane. Our data suggests that BPs not only form ordered superstructures, but also rather homogeneously dispersed π-packed filaments within the lipid gel phase, thus reducing the corrugation of large vesicles.

Comparison between the dynamics of lipid/gramicidin A systems in the lamellar and hexagonal phases: a solid-state 13C NMR study

We have investigated the effect of gramicidin A on the dynamics of two model membranes: dimyristoylphosphatidylcholine (DMPC) in the lamellar phase at a lipid-to-peptide molar ratio of 10:1 and dioleoylphosphatidylcholine (DOPC) in the hexagonal HII phase at a lipid-to-peptide molar ratio of 5:1. Natural abundance 13C nuclear magnetic resonance (NMR) spectroscopy was used in combination with magic angle spinning to increase the spectral resolution, therefore allowing the different regions of the lipid bilayers to be investigated from the same spectra. 31P NMR was also used to detect and confirm the formation of the DOPC HII phase in the presence of gramicidin A. In order to examine the effect of gramicidin A on both the fast and slow motions of DMPC and DOPC, the 1H spin-lattice relaxation times in the laboratory frame (HT1) as well as the 1H spin-lattice relaxation times in the rotating frame (HT1rho) were calculated for each resolved protonated lipid resonance in the 13C spectra. For both DMPC and DOPC, we found that the presence of gramicidin A does not significantly affect the fast motions of the lipid acyl chains but increases slightly the fast motions of the polar head group. However, the HT1rho are significantly decreased, this effect being more pronounced for DOPC most likely due to a decrease in the rate of the lipid lateral diffusion.

Membrane protein structure: the contribution and potential of novel solid state NMR approaches

Alternative methods for describing molecular detail for large integral membrane proteins are required in the absence of routine crystallographic approaches. Novel solid state NMR methods, devised for the study of large molecular assemblies, are now finding applications in biological systems, including integral membrane proteins. Wild-type and genetically engineered proteins can be investigated and detailed information about side chains, prosthetic groups, ligands (e.g. drugs) and binding sites can be deduced. The molecular structure and dynamics of selected parts of the proteins are accessible by a range of different solid state NMR approaches. Inter- and intra-atomic distances can be determined rather accurately (within ångströms) and the orientation of molecular bonds (within 2 degrees) can be measured in ideal cases. Here, a brief description of the methods is given and then some specific examples described with an indication of the future potential for the approaches in studying membrane proteins. It is anticipated that this emerging NMR methodology will be more widely used in the future, not only for resolving local structure, but also for more expansive descriptions of membrane protein structure at atomic resolution.

Restatement of order parameters in biomembranes: calculation of C-C bond order parameters from C-D quadrupolar splittings

An expression for the C-C bond order parameter, SCC, of membrane hydrocarbon chains has been derived from the observed C-D bond order parameters. It allows calculation of the probability of each of the C-C bond rotamers and, consequently, the number of gauche defects per chain as well as their projected average length onto the bilayer normal, thus affording the calculation of accurate hydrophobic bilayer thicknesses. The effect of temperature has been studied on dilauroyl-, dimyristoyl-, and dipalmitoylphosphatidylcholine (DLPC, DMPC, DPPC) membranes, as has the effect of cholesterol on DMPC. The salient results are as follows: 1) an odd-even effect is observed for the SCC versus carbon position, k, whose amplitude increases with temperature; 2) calculation of SCC, from nonequivalent deuterons on the sn-2 chain of lipids, SCC2, leads to negative values, indicating the tendency for the C1-C2 bond to be oriented parallel to the bilayer surface; this bond becomes more parallel to the surface as the temperature increases or when cholesterol is added; 3) calculation on the sn-2 chain length can be performed from C1 to Cn, where n is the number of carbon atoms in the chain, and leads to 10.4, 12.2, and 13.8 A for DLPC, DMPC, and DPPC close to the transition temperature, TC, of each of the systems and to 9.4, 10.9, and 12.6 for T-TC = 30-40 degrees C, respectively; 4) separation of intra- and intermolecular motions allows quantitation of the number of gauche defects per chain, which is equal to 1.9, 2.7, and 3.5 for DLPC, DMPC, and DPPC near TC and to 2.7, 3.5, and 4.4 at T-TC = 30-40 degrees C, respectively. Finally, the validity of our model is discussed and compared with previously published models.

2H-NMR investigation of DMPC/glycophorin bilayers

Deuterium nuclear magnetic resonance spectroscopy was used to investigate the phase equilibria, and the temperature and concentration dependences of the phospholipid hydrocarbon chain order, of mixtures of glycophorin in dimyristoylphosphatidyl-choline. In the fluid phase it is found that the protein has only a slight effect on the first moment of the 2H spectrum, which for perdeuterated chains is a direct measure of the average chain orientational order. However, analysis of the rate of change of the first moment with respect to protein concentration, at different temperatures within the fluid phase, shows that at a molar protein concentration of about 0.0295 +/- 0.01, the lipid chain order (or M1) is essentially independent of temperature. At this concentration the chain order is determined by the lipid's interaction with the protein and one can conclude that about 34 (+/- 12) lipids are required to solvate the protein. At higher lipid concentrations these lipids are freely exchanging, on the NMR time scale, with the other lipids in the bilayer. At glycophorin concentrations below about 1 mol% there is a two-phase coexistence region at temperatures below the pure lipid's chain melting transition. The boundary between the fluid phase and this two-phase region curves downwards (is concave downwards), whereas the boundary between the two-phase region and the gel phase, while naturally occurring at lower temperatures than the upper boundary, is concave upwards. As a consequence the protein partitions preferentially into the fluid phase. This behaviour is similar to that observed in a number of other protein/lipid and peptide/lipid mixtures where it was suggested that those systems may have been close to a critical mixing point and some characteristics of a continuous phase change were noted. Indeed, at glycophorin concentrations near and above 1 mol% there are indications that the phase behaviour becomes more complex, suggesting the presence of significant protein/protein interactions and that this system may be close to a critical point.

Effect of vesicle composition and curvature on the dissociation of phosphatidic acid in small unilamellar vesicles--a 31P-NMR study

Sonicated small unilamellar vesicles (SUVs) containing phosphatidic acid (PA) give two PA 31P-NMR resonances corresponding to PA molecules in the inner and outer leaflets of the bilayer. This NMR differentiation between the two monolayers is not due to a pH gradient across the membrane but instead reflects differential packing in the inner and outer leaflets imposed by the highly curved SUV surface. The apparent pKa of the outer-leaflet PA increases with decreasing surface curvature and with increasing PA content. The estimated relationship between the apparent pKa of the outer-leaflet PA headgroup and vesicle curvature may provide a qualitative probe for effects related to surface curvature in these model-membrane systems.

Chronic ethanol intoxication induces adaptive changes at the membrane protein/lipid interface

Modifications were found to occur at the membrane protein/lipid interface of liver microsomes in animals that had been subjected to chronic ethanol ingestion. The effects were revealed by probing this region with 1,6-diphenyl-1,3,5-hexatriene (DPH), trimethylammonium-DPH (TMA-DPH) and DPH attached to the sn-2 chain of phosphatidylcholine (1-palmitoyl-2-[[2-[4-(6-phenyl-trans-1,3,5-hexatrienyl) phenyl]ethyl]carbonyl]-3-sn-phosphatidylcholine, DPH-PC). In intact membranes, it was found that the decay of the excited state was heterogeneous, this being modeled by fitting the data to a fluorescence lifetime distribution. The full-width of the distribution at half-maximum, which relates to the degree of excited state environmental heterogeneity, increased for each fluorophore, as a result of chronic ethanol treatment. For TMA-DPH and DPH the excited state heterogeneity could have arisen from, (i) the protein/lipid interface and (ii) varied degrees of water penetration into the lipid, due to the ability of these fluorophores to sample along the bilayer normal. By contrast, the DPH in DPH-PC, due to its tethering, was only able to sample the heterogeneity at the protein/lipid interface, as confirmed by a homogeneous decay in vesicles of microsomal lipid extracts. The increased degree of DPH-PC fluorescence decay heterogeneity in microsomes from chronic ethanol-treated animals as compared to controls, was found to persist in vesicles of extracted lipids, when apocytochrome C was included in the vesicle preparations as a model protein. This effectively eliminated a protein modification from being responsible and indicated that a chronic-ethanol induced alteration in the lipids was being expressed in the form of a physico-chemical modification at the protein/lipid interface. The degree of DPH-PC environmental heterogeneity was also directly increased by ethanol, however, membranes from chronic ethanol-treated animals were resistant to this effect, showing that the phenomenon of 'membrane tolerance' extends to the membrane protein/lipid interface.

An electron spin resonance study of interactions between gramicidin A' and phosphatidylcholine bilayers

The model of microscopic order and macroscopic disorder was used to stimulate electron spin resonance spectra of spin-labeled lipids, 5-PC, 10-PC, and 16-PC in multilamellar vesicles of dipalmitoylphosphatidylcholine (DPPC) containing gramicidin A' (GA) at temperatures above the gel-to-liquid crystal transition of DPPC. The simulations show that at a lower concentration of GA (i.e., molar ratios of DPPC/GA greater than 3), GA has only a slight effect on the acyl chain dynamics. The rotational diffusion rate around the axis parallel to the long hydrocarbon chain remains unchanged or increases slightly, while the rate around the perpendicular axes decreases slightly. These spectra from DPPC/GA mixtures could only be fit successfully with two or more components consistent with the well-known concept of "boundary lipids," that is, the peptide induces structural inhomogeneity in lipid bilayers. However, the spectra were significantly better fit with additional components that exhibit increased local ordering, implying decreased amplitude of rotational motion, rather than immobilized components with sharply a reduced rotational rate. The largest relative effects occur at the end of the acyl chains, where the average local order parameter St of 16-PC increases from 0.06 for pure lipid to 0.66 for 1:1 DPPC/GA. The inhomogeneity in ordering in DPPC bilayers due to GA decreases with increasing temperature. The hyperfine tensor component Azz increases for 10-PC and 16-PC when GA is incorporated into DPPC bilayers, indicating that water has deeply penetrated into the DPPC bilayers. Simulations of published electron spin resonance spectra of 14-PC in dimyristoylphosphatidylcholine/cytochrome oxidase complexes were also better fit by additional components that were more ordered, rather than immobilized. The average local order parameter in this case is found to increase from 0.11 for pure dimyristoylphosphatidylcholine to 0.61 for a lipid/protein ratio of 50. These spectra and their simulations are similar to the results obtained with 16-PC in the DPPC/GA mixtures. The relevance to studies of lipid-protein interactions for other proteins is briefly discussed.

Spin label saturation transfer EPR determinations of the stoichiometry and selectivity of lipid-protein interactions in the gel phase

Lipid-protein interactions with the myelin proteolipid protein incorporated in the gel phase of dimyristoylphosphatidylcholine bilayers have been studied by saturation transfer EPR spectroscopy of spin-labelled phospholipids. The integrated intensities of the saturation transfer EPR spectra from spin-labelled phosphatidylcholine are linearly dependent on the protein/lipid ratio, and correspond to a fixed stoichiometry of approximately 11 lipids per monomer associated with the protein in the gel phase. The normalized saturation transfer intensities of spin-labelled phosphatidic acid, on the other hand, display a non-linear dependence on the protein/lipid ratio that can be described well by a selectivity for interaction with the protein in the gel phase with an average association constant relative to phosphatidylcholine of approx. 5.2. These values for the stoichiometry and selectivity of lipid-protein interaction in the lipid gel phase obtained from saturation transfer EPR spectroscopy are comparable to those found previously in fluid phase lipids by conventional EPR spectroscopy.

Formation of non-bilayer structures induced by M13 coat protein depends on the conformation of the protein

A comparison is made of the interaction of the coat protein of bacteriophage M13 in a predominant alpha-helix conformation and in a predominant beta-sheet conformation. To perform a systematic study of the interaction between the protein in these two different forms of the surrounding lipid matrix, NMR spectra of 2H-nuclei of specific labelled phospholipid systems are measured. In addition 31P-NMR is employed to provide information about the morphological structure adopted by the reconstituted lipid/protein systems. From the 2H-NMR studies on specific headgroup and chain deuterium labelled phospholipids it is found that the protein in the predominant beta-sheet conformation causes a fraction of lipids to be trapped. By combining the results from the headgroup and acyl chains of the phospholipids, it is concluded that the trapped lipids are arranged in a non-bilayer structure, probably caused by a misfitting of the hydrophobic core of the protein and the membrane bilayer. The protein in the predominant alpha-helix conformation perfectly fits in the lipid bilayer and has only minor influences on the surrounding lipid matrix. A new model is proposed to explain the presence of the trapped lipids in the lipid/protein systems.

A NMR investigation on the interactions of the alpha-oligomeric form of the M13 coat protein with lipids, which mimic the Escherichia coli inner membrane

The interaction of the M13 bacteriophage major coat protein in the alpha-oligomeric form with specifically deuterated phospholipid headgroups which mimic the Escherichia coli inner membrane, has been studied using NMR methods. As can be seen from the deuterium NMR spectra obtained with headgroup trimethyl deuterated DOPC, the coat protein in the alpha-oligomeric form does not give rise to trapped lipids as observed with M13 coat protein in the beta-polymeric form (Van Gorkom et al. (1990) Biochemistry 29, 3828-3834). The quadrupolar splittings of the alpha headgroup methylene deuterons of deuterated phosphatidylcholine and phosphatidylethanolamine decrease, whereas the quadrupolar splittings of the beta headgroup methylene deuterons of the two lipids increase with increasing protein content. All deuterated segments in the phosphatidylglycerol headgroup show the same relative decrease of the NMR quadrupolar splittings. These results are interpreted in terms of a change in torsion angles of the methylene groups, induced by positive charges, probably lysine residues of the protein at the membrane surface. For all lipid bilayer compositions studied the head-group perturbations are similar. It is concluded that there is no strong specific interaction between one of the lipid types examined and the M13 coat protein. From the spin-spin (T2e) relaxation time and spin-lattice (T1z) relaxation time of all deuterated lipids it is concluded that at the bilayer surface only slow motions are affected by the M13 coat protein.

Effect of cholesterol and lanosterol on the structure and dynamics of the cell membrane of Mycoplasma capricolum. Deuterium nuclear magnetic resonance study

Deuterium nuclear magnetic resonance (NMR) techniques were employed to study the effect of sterols on the composition and dynamics of the membrane lipids of Mycoplasma capricolum, a natural fatty acid auxotroph that requires sterols for growth. The membrane lipids of cells grown in modified Edwards medium supplemented with cholesterol, oleic acid (OA), and palmitic acid (PA) were composed primarily of phosphatidylglycerol (PG) (60%) and cardiolipin (CL) (35%). The incorporation of cholesterol and the cellular OA/PA ratio increased nonlinearly with increases in exogenous cholesterol level, whereas the levels of phospholipid increased only slightly. At the growth temperature, 37 degrees C, the residual deuterium quadrupole splittings were found to be 43-46 kHz for cells grown with (7,7,8,8-2H4) PA and 1.25 micrograms/ml (30 mol%) to 10 micrograms/ml (50 mol%) cholesterol, respectively, similar to that found in the cholesterol/lecithin binary dispersions of similar cholesterol contents. Deuterium T2e of these samples were found to be 170 +/- 10 microseconds and were independent of cellular cholesterol content. In comparison, T2e of the corresponding lipid extracts were longer (320-420 microseconds) and dependent on cholesterol content. Thus, lipid-protein interactions in the cell membrane is the dominant mechanism responsible for the reduced T2e. At lower temperatures, spectra indicative of the coexistence of gel and liquid-crystalline states were observed for cells having low cholesterol levels. For both cell membrane and membrane lipid extract containing 50 mol% cholesterol, T2e was found to be constant at the temperature range from 15 to 40 degrees C. On the other hand, T2e of cell membrane containing 30 mol% cholesterol decreased linearly at 3.2 microseconds/degrees C. T2e of the corresponding lipid extract showed much stronger temperature variation. Cells containing 39 mol% lanosterol were found to have a quadrupole splitting of 39 kHz, broader than that of the cholesterol-free lecithin dispersion (less than 30 kHz) but less than that of cell membrane containing 30 mol% cholesterol (43 kHz). T2e of the lanosterol sample was found to be 130 +/- 10 microseconds which decreased linearly at a slope similar to that observed for the low cholesterol sample. Therefore, although lanosterol appeared to be capable of modulating cell membrane physical properties it is less effective than cholesterol. When growth rates were correlated with NMR parameters, we found that the membranes of faster growing cells were also more ordered. In contrast, the T2e of the cells of M. capricolum seemed to be maintained at a relatively constant value around 170 microseconds.

Phospholipid phase transitions as revealed by NMR

Aqueous dispersions of phospholipids can adopt a range of polymorphic phases which include bilayer and non-bilayer forms. Within the bilayer form, laterally separated phases may be induced as a result of surface electrostatic associations, thermotropic behaviour, lipid-protein interactions or because of molecular mismatch between chemically distinct phospholipids. Nuclear magnetic resonance (NMR) methods, designed to exploit the properties of either indigenous nuclei or isotopic labels introduced specifically into a phospholipid, can be used in some cases to describe the molecular properties and behaviour of phospholipids in both macroscopically distinct phases and in molecularly distinct phases within the same polymorphic state. If the molecular motion of phospholipids in co-existing phases is sufficiently different, NMR methods can, in principle, give estimates of the life-time of the phases and the rate of molecular exchange between the phases.

Interaction between perdeuterated dimyristoylphosphatidylcholine and low molecular weight pulmonary surfactant protein SP-C

A low molecular weight hydrophobic protein was isolated from porcine lung lavage fluid using silicic acid and Sephadex LH-20 chromatography. The protein migrated with an apparent molecular weight of 5000-6000 on SDS-PAGE under reducing and nonreducing conditions. Gels run under reducing conditions also showed a minor band migrating with a molecular weight of 12,000. Amino acid compositional analysis and sequencing data suggest that this protein preparation contains intact surfactant protein SP-C and about 30% of truncated SP-C (N-terminal leucine absent). The surfactant protein was combined with perdeuterated dimyristoylphosphatidylcholine (DMPC-d54) in multilamellar vesicles. The protein enhanced the rate of adsorption of the lipid at air-water interfaces. The ability of the protein to alter normal lipid organization was examined by using high-sensitivity differential scanning calorimetry (DSC) and 2H nuclear magnetic resonance spectroscopy (2H NMR). The calorimetric measurements indicated that the protein caused a decrease in the temperature maximum (Tm) and a broadening of the phase transition. At a protein concentration of 8% (w/w), the enthalpy change of transition was reduced to 4.4 kcal/mol compared to 6.3 kcal/mol determined for the pure lipid. NMR spectral moment studies indicated that protein had no effect on lipid chain order in the liquid-crystal phase but reduced orientational order in the gel phase. Two-phase coexistence in the presence of protein was observed over a small temperature range below the pure lipid transition temperature. Spin-lattice relaxation times (T1) were not substantially affected by the protein. Transverse relaxation time (T2e) studies suggest that the protein influences slow lipid motions.

Transverse nuclear spin relaxation in phosphatidylcholine bilayers containing gramicidin

Deuterium nuclear magnetic resonance has been used to study transverse relaxation in samples of 1,2-dimyristoyl-sn-glycero-3-phosphocholine, perdeuterated and specifically deuterated at the alpha position of the chains, containing the polypeptide gramicidin at concentrations of 0, 1, and 4 mol%. For 4 mol% gramicidin, the bilayer is thought to undergo a continuous phase change rather than a phase transition proceeding via two phase coexistence. Information is obtained regarding lipid dynamics in the continuous phase change region of the phase diagram. In the presence of gramicidin, the transverse relaxation time measured by the quadrupole echo technique, T2e, passes through a minimum in the gel phase. The gramicidin concentration dependence of T2e suggests that the polypeptide reduces the temperature sensitivity of the correlation time responsible for the minimum. The polypeptide also increases the sensitivity of the first spectral moment, M1, to the quadrupole echo pulse separation. This behavior is attributed to a polypeptide-induced enhancement of the spread in T2e along the acyl chains. Quadrupole Carr-Purcell-Meiboom-Gill experiments are used to separate contributions to the observed behavior from fast and slow motions.

Effect of proteins on fluorophore lifetime heterogeneity in lipid bilayers

The effect of three different membrane proteins on the fluorescence lifetime heterogeneity of 1,6-diphenyl-1,3,5-hexatriene (DPH) in phospholipid vesicle systems was investigated. For large unilamellar vesicles of dimyristoylphosphatidylcholine (DMPC) and 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) at 37 degrees C, the fluorescence decay was essentially monoexponential (8.6 and 8.2 ns, respectively) except for a minor component typical of DPH. For gramicidin D reconstituted into DMPC vesicles at a protein/lipid molar ratio of 1/7, the most appropriate analysis of the data was found to be in the form of a bimodal Lorentzian distribution. Centers of the major lifetime components were almost identical with those recovered for vesicles without proteins, while broad distributional widths of some 4.0 ns were recovered. Variation of the protein/lipid molar ratio in sonicated POPC vesicles revealed an abrupt increase in distributional width at ratios approximating 1/15-1/20, which leveled off at about 2.5 ns. For bacteriorhodopsin in DMPC vesicles and cytochrome b5 in POPC, the most appropriate analysis of the data was again found to be in the form of a bimodal Lorentzian also with broad distributional widths in the major component. Lifetime centers were decreased for these proteins due to fluorescence energy transfer to the retinal of the bacteriorhodopsin and heme of the cytochrome b5. Fluorescence energy transfer is distance dependent, and since a range of donor-acceptor distances would be expected in a membrane, lifetime distributions should therefore be recovered independently of other effects for proteins possessing acceptor chromophores.(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of phospholipids on the thermal stability of microsomal UDP-glucuronosyltransferase

The GT2P isoform of microsomal UDP-glucuronosyltransferase from pig liver is a lipid-dependent enzyme. The data in the present work indicate that, in addition to regulation of activity, the thermal stability of the enzyme also is modulated by the acyl chain composition of phosphatidylcholines (PC) used to reconstitute the activity of pure enzyme. There was a reversible, temperature-dependent change in the state of the pure enzyme to an inactive form with onset at T greater than 38 degrees C, depending on the environment of the enzyme. The midpoint for the transition shifted from 39.8 degrees C for enzyme in a bilayer of distearoylphosphatidylcholine (DSPC) to 47.5 degrees C for enzyme in a bilayer of 1-stearoyl-2-oleoylphosphatidylcholine (SOPC). For all lipids, the transition from a catalytically active to an inactive form of the enzyme was associated with large compensating changes in H and S. Lipid-induced stabilization of the active form of UDP-glucuronosyltransferase at T greater than 37 degrees C was associated with decreases in delta H and delta S, but the decreases in delta S were larger, indicating that lipid-induced stabilization of the active form of the enzyme was entropic. The transition between the active and inactive forms of the enzyme was too rapid in either direction to measure in a standard spectrophotometer. In addition to reversible inactivation of the enzyme, there was a slower irreversible, temperature-dependent inactivation. The rate of this process depended on the acyl chains of the phosphocholines interacting with the enzyme. However, there was no obvious correlation between the structures of lipids that stabilized the different inactivation reactions.

Selective detection of the rotational dynamics of the protein-associated lipid hydrocarbon chains in sarcoplasmic reticulum membranes

We have developed a saturation transfer EPR (ST-EPR) method to measure selectively the rotational dynamics of those lipids that are motionally restricted by integral membrane proteins and have applied this methodology to measure lipid-protein interactions in native sarcoplasmic reticulum (SR) membranes. This analysis involves the measurement of spectral saturation using a series of six stearic acid spin labels that are labeled with a nitroxide at different carbon atom positions. A large amount of spectral saturation is observed for spin labels in native SR membranes, but not for spin labels in dispersions of extracted SR lipids, implying that the motional properties of those lipids interacting with the Ca-ATPase, i.e., the boundary or annular lipid, can be directly measured without the need for spectral subtraction procedures. A comparison of the motional properties of the restricted lipid, measured by ST-EPR, with those measured by digital subtraction of conventional EPR spectra qualitatively agree, for in both cases the Ca-ATPase restricts the rotational mobility of a population of lipids, whose rotational mobility increases as the nitroxide is positioned toward the center of the bilayer. However, the ability of ST-EPR to directly measure the motionally restricted lipid in a model-independent means provides the greater precision necessary to measure small changes in the rotational dynamics of the lipid at the protein-lipid interface, providing a valuable tool in clarifying the relationship between the physical nature of the protein-lipid interface and membrane function.

Dynamics and orientation of glycolipid headgroups by 2H-NMR: gentiobiose

Deuterium nuclear magnetic resonance has been used to investigate the dynamics and determine the orientation of the headgroup of the glycolipid 1,2-di-O-tetradecyl-3-O-(6-O-beta-D-glucopyranosyl-beta-D-glucopyranosyl )-sn- glycerol (beta-DTDGL), in aqueous multilamellar dispersions. In addition, its anomeric analog, having an alpha glucose-glycerol linkage, was prepared and examined. The lipids were labelled with deuterium at specific positions in the disaccharide moiety. Analysis of the deuterium quadrupolar splittings for the first glucose ring (glycerol-linked) gave segmental order parameters of 0.43 and 0.35 for the beta and alpha isomers, respectively. Both isomers had similar orientations of the sugar ring relative to the bilayer surface, as determined for lipid in the liquid-crystalline phase. 2H-NMR results for the lipid labelled at C-6' are consistent with a single conformation about the C-5'-C-6' bond of the first glucose residue, with a dihedral angle (O-5'-C-5'-C-6'-O-6') of -17 degrees. The results obtained for the second sugar ring suggest that two conformers may be present, which are in slow exchange on the 2H-NMR timescale. Measurements of longitudinal relaxation times, T1z, gave similar values for both sugar moieties in the headgroup, suggesting that the disaccharide does not exhibit the flexibility expected about the 1----6 linkage. Since T1z for 2H in these compounds decreases with increasing temperature and increases with magnetic field strength, the motion(s) dominating relaxation is in the long-correlation-time regime [omega 0 tau c)2 greater than 1). Thus, the gentiobiosyl headgroup undergoes the slowest motion of the glycolipid headgroups studied to date.

Deuterium nuclear magnetic resonance investigation of bacteriophage M13 coat protein in dimyristoylphosphatidylcholine liposomes using palmitic acid as a probe

The effect of incorporation of various amounts of M13 bacteriophage coat protein on the bilayer order and acyl chain motion in dimyristoylphosphatidylcholine (DMPC) liposomes has been investigated using deuterium NMR of specifically deuterated palmitic acid as a bilayer probe, phosphorus NMR and additional spin-label electron spin resonance (ESR). The secondary structure of the M13 coat protein in these bilayers was determined from circular dichroism spectra. Phosphorus NMR spectra of the mixed liposomes are characteristic for DMPC organized in bilayers, also after incorporation of various levels of M13 protein. Circular dichroism spectra of the coat protein indicate that the protein conformation is predominantly a beta-structure (more than 75%). Various incorporation levels of M13 coat protein do not affect the order of the deuterium-labelled positions along the acyl chain at the carbon-2, 9 and 16 positions. In contrast, the spin-spin relaxation times decrease at higher protein levels, especially at the carbon-16 position. The spin-label ESR spectra of the same system using 14-doxylstearic acid as a label show a second, motionally restricted component, that is not observed by deuterium NMR. The NMR and ESR results are consistent with a model in which the fatty acid molecules are in a fast two-site exchange (at a rate of approx. 10(7) Hz) between the sites in the bulk of the lipid bilayer and the motionally restricted sites on the coat protein.

Differential scanning calorimetry and 2H NMR studies of the phase behavior of gramicidin-phosphatidylcholine mixtures

The extents of two-phase coexistence in the phase diagrams of mixtures of gramicidin with 1,2-bis(perdeuteriopalmitoyl)-sn-glycero-3-phosphocholine (DPPC-d62) and with 1,2-bis(perdeuteriomyristoyl)-sn-glycero-3-phosphocholine (DMPC-d54) mixtures have been explored with differential scanning calorimetry (DSC) and deuterium nuclear magnetic resonance (2H NMR). For both systems, increased gramicidin content causes a decrease in transition enthalpy and a broadening of the peak in excess heat capacity at the transition. In DMPC-d54-based mixtures, the broadening is roughly symmetric about the pure lipid transition temperature. Addition of gramicidin to DPPC-d62 extends the excess heat capacity peak on the low-temperature side, resulting in a slightly asymmetric scan. Deuterium NMR spectra showing a superposition of gel and liquid-crystalline components, observed for both mixtures, indicate the presence of two-phase coexistence. For the DPPC-d62-based mixtures, two-phase coexistence is restricted to an approximately 2 degrees C temperature range below the pure transition temperature. For DMPC-d54-based mixtures, the region of two-phase coexistence is even narrower. For both mixtures, beyond a gramicidin mole fraction of 2%, distinct gel and liquid-crystal contributions to the spectra cannot be distinguished. Along with the broad featureless nature of the DSC scan in this region, this is taken to indicate that the transition has been replaced by a continuous phase change. These results are consistent with the existence of a closed two-phase region having a critical concentration of gramicidin below 2 mol%.

Nuclear magnetic resonance methods to characterize lipid-protein interactions at membrane surfaces

Specific molecular interactions that determine many of the functions of a biomembrane have a high probability of occurring at the surface of that membrane. However, unlike their hydrophobic core, the polar-apolar interface of biomembranes has been somewhat neglected experimentally. Reasons for this are that the chemical heterogeneity encountered makes a simple description difficult and that probing the membrane surface often involves a perturbation of those very interactions being studied. Classical methods for obtaining structural information about biomolecules, including X-ray diffraction, electron microscopy, and more recently high-resolution 2D nuclear magnetic resonance techniques are inappropriate for all but the simplest of membrane problems. In an effort to throw light on how membrane surfaces are organized, both architecturally and dynamically, protons in lipids and proteins have been selectively replaced by deuterons and the resultant deuterium NMR spectrum analyzed to give structural and dynamic information about the molecular associations between a range of membrane components. In principle, lipids, proteins, and oligosaccharides can be studied by this method and the information gained related to biochemical integrity and function. With one or two notable exceptions, the majority of the studies reported so far have been on model systems. A comprehensive review of the literature will not be presented here. However, protein-lipid molecular specificity in membranes, peptide-induced lateral separation, and the ionization behavior of deuterated phospholipids and peripheral proteins will all be demonstrated predominantly using deuterium NMR methods. Some suggestions for future work are also presented.

Dynamics of the phosphate group in phospholipid bilayers. A 31P nuclear relaxation time study

The spin-lattice relaxation time of the 31P nucleus in the phosphate group of egg yolk phosphatidylcholine multilamellar dispersions has been investigated at four resonant frequencies (38.9, 81.0, 108.9, and 145.7 MHz) in the temperature range from -30 degrees to 60 degrees C. The observed frequency dependence of the relaxation indicates that both dipolar relaxation and relaxation due to anisotropic chemical shielding are significant mechanisms. The experimental data have thus been modeled assuming both mechanisms and the analysis has allowed the contribution of each to the relaxation to be determined along with the correlation time for the molecular reorientation as a function of temperature. Dipolar relaxation was found to dominate at low nuclear magnetic resonance frequencies while at high frequencies the anisotropic chemical shift dominates. The correlation time of the phosphate group is on the order of 10(-9) s at 60 degrees C and increases to approximately 10(-7) s at -30 degrees C. It is observed that the freezing of the buffer which occurs at approximately -8 degrees C has a significant effect on the phosphate group reorientation. This effect of the freezing is to change the activation energy for the phosphate group reorientation from 16.9 KJ/mol above -8 degrees C to 32.5 KJ/mol below -8 degrees C.

Molecular details of melittin-induced lysis of phospholipid membranes as revealed by deuterium and phosphorus NMR

Solid-state deuterium and phosphorus-31 nuclear magnetic resonance (2H and 31P NMR) studies of deuterium-enriched phosphatidylcholine [( 3',3'-2H2]DPPC, [sn-2-2H31]DPPC) and ditetradecylphosphatidylglycerol (DMPG-diether), as water dispersions, were undertaken to investigate the action of melittin on zwitterionic and negatively charged membrane phospholipids. When the lipid-to-protein ratio (Ri) is greater than or equal to 20, the 2H and 31P NMR spectral features indicate that the system is constituted by large bilayer structures of several thousand angstrom curvature radius, at T greater than Tc (Tc, temperature of "gel-to-liquid crystal" phase transition of pure lipid dispersions). At T approximately Tc, a detailed analysis of the lipid chain ordering shows that melittin induces a slight disordering of the "plateau" positions concomitantly with a substantial ordering of positions near the bilayer center. At T much greater than Tc, an apparent general chain disordering is observed. These findings suggest that melittin is in contact with the acyl chain segments and that its position within the bilayer may depend on the temperature. On a cooling down below Tc, for Ri greater than 20, two-phase spectra are observed, i.e., narrow single resonances superimposed on gel-type phosphorus and deuterium powder patterns. These narrow resonances are characteristic of small structures (vesicles, micelles, ... of a few hundred angstrom curvature radius) undergoing fast isotropic reorientation, which averages to zero both the quadrupolar and chemical shift anisotropy interactions. On an increase of the temperature above Tc, the NMR spectra indicate that the system returns reversibly to large bilayer structures.(ABSTRACT TRUNCATED AT 250 WORDS)

Monte Carlo calculations of order parameter profiles in models of lipid-protein interactions in bilayers

The Monte Carlo method has been utilized to calculate lipid chain order parameters in model monomolecular layers (half-bilayers) containing several different model polypeptides. The systems all consist of a periodic array of identical cells, each containing 35 hydrocarbon chains and 1 "perturbant" (a small model polypeptide or protein). The lipid chains are each 10 CH2 subunits long, have one end constrained to lie in the bilayer plane, and interact via van der Waals forces between all subunits. The chains also interact with the perturbant via van der Waals forces. With standard Monte Carlo procedures order parameter profiles are calculated for chains that are close to the perturbant and for the nonneighboring chains. In order to examine a wide range of possibilities, several different model polypeptides are considered: (i) a rigid smooth cylinder, (ii) a cylinder with identical side chains at alpha-helical positions, (iii) a cylinder with nonidentical side chains at alpha-helical positions, and (iv) a cylinder identical with (ii) but which only extends about halfway through the monolayer. Although results differ for the different systems studied, in all cases only slight conformational differences between the bulk chains and the chains that are nearest the perturbants are found, and it is not possible to characterize the boundary chains as "more ordered" or "less ordered" than the nonboundary chains.

A detailed analysis of the motions of cholesterol in biological membranes by 2H-NMR relaxation

Spin-lattice relaxation, T1z, measurements of [2,2,3,4,4,6-2H6]cholesterol in model membranes of DMPC were performed as a function of temperature, Larmor frequency and position of labelling in the fused ring system. The results are interpreted according to a hierarchy of motions, such that motion i of correlation time tau i reduces the residual ordering set, characterizing motions i-1, i-2, etc..., by the amount Si = d(2)00(beta i), where beta i is the angle between the axes of motional averaging of motions i and i-1, respectively and d(2)00 is the Wigner rotation matrix element. The appearance of minima in the temperature dependence of T1z for cholesterol, at 46.1 MHz and 30.7 MHz, and the scaling of these T1z (min) according to the orientation of each individual C-2H bond with respect to the axis of motional averaging of cholesterol, allows assignment of the sterol axial rotation to the second fastest motion, characterized by a correlation time of 3.2 X 10(-9) s at 25 degrees C and an activation energy of 32 +/- 5 kJ X mole-1. The fastest motion of cholesterol in DMPC could be a very rapid libration, 'wobbling', which does not contribute significantly to the T1z relaxation of cholesterol at physiological temperatures and Larmor frequencies smaller than 50 MHz, but does reduce the ordering of the cholesterol molecule in DMPC from S0 = 1 to S1 = 0.8, at 25 degrees C.

Studies on the interaction of human erythrocyte band 3 with membrane lipids using deuterium nuclear magnetic resonance and differential scanning calorimetry

Human erythrocyte band 3, reconstituted into large unilamellar phospholipid vesicles, has been used as a model system for studying the interactions between membrane lipids and large transmembrane glycoproteins. Both 2H-nuclear magnetic resonance (2H-NMR) and differential scanning calorimetric techniques have been used to probe dimyristoylphosphatidylcholine-band 3 interactions over the temperature range 4-32 degrees C. Analysis of 2H-NMR spectra allowed the assignment of liquid crystal, gel phase and two-phase regions for several protein/lipid mole fractions in the range (1-20) X 10(-4). Sample size was limited by the amount of available glycoprotein and this precluded exact determination of the phase boundaries for this system. The sharp discontinuity in the spectral first moment, M1, seen at the phase transition of the pure phospholipid is progressively diminished by addition of protein, and at the highest protein concentration the first moment varies smoothly between the two phases. For T greater than 26 degrees C or less than 16 degrees C, the moments are relatively insensitive to protein concentration, while between 20 and 26 degrees C the moments increase with protein concentration up to the boundary of the two-phase region. Beyond this boundary, they remain constant or decrease slightly with increasing amount of protein. A preliminary phase diagram for band 3 in this lipid system is presented, based on 2H-NMR data. Differential scanning calorimetry (DSC) showed that addition of glycoprotein dramatically alters the scan shape and tends to extend the coexistence of two phases to higher temperatures.

The influence of membrane proteins on lipid dynamics

The application of electron paramagnetic resonance (EPR), and nuclear magnetic resonance (NMR) to the study of phospholipid dynamics in membranes is discussed. Using these complementary spectroscopic techniques it is possible to investigate the dynamics of lipids in membranes over a time scale range of from 10(-10) to 1 s. A rather detailed, quantitative description of phospholipid dynamics in pure lipid/water bilayer dispersions has emerged. For example, the correlation time for phosphate group reorientation has been shown to be of the order of 10(-9) s. Chain dynamics can be modelled in terms of three basic types of motion: reorientation about the long axis, fluctuation of the long axis with respect to the bilayer normal, and gauche-trans isomerization about C-C bonds. In the fluid phase, all of these chain motions are in the fast limit on the NMR time scale, but only the gauche-trans isomerization is fast on the EPR time scale. In the gel phase, all of these motions are in the intermediate time scale regime for NMR. While a similarly detailed description of the influence of protein on lipid dynamics has not yet been obtained, these techniques have demonstrated their capability to perform that task. The limited data available suggest that the major effect of protein on lipid dynamics is to increase the relative importance of motions at lower frequency. This is most clearly evident as a slight increase in the correlation time for phosphate group reorientation. The strongest evidence for slower motion of the hydrocarbon chains is from NMR relaxation time and line width measurements. The interpretation of changes in lipid dynamics in terms of protein/lipid interactions will require further studies of protein/lipid phase equilibria as well as molecular dynamics.

Simultaneous observation of order and dynamics at several defined positions in a single acyl chain using 2H NMR of single acyl chain perdeuterated phosphatidylcholines

Deuterium nuclear magnetic resonance (2H NMR) spectra from aqueous dispersions of phosphatidylcholines in which perdeuterated palmitic acid is esterified at the sn-1 position have several very useful features. The powder spectra show six well-resolved 90 degree edges which correspond to the six positions closest to the methyl end of the acyl chain. The spectral overlap inherent in the multiple powder pattern line shape of these dispersions can be removed by using a "dePaking" procedure [Bloom, M., Davis, J.H., & Mackay, A. (1981) Chem. Phys. Lett. 80, 198-202] which calculates the spectra that would result if the lipid bilayers were oriented in the magnetic field. This procedure produces six well-resolved doublets whose NMR properties can be observed without interference from the resonances of other labeled positions. The presence of a single double bond in the sn-2 chain increases the order of the saturated 16:0 sn-1 chain at every position in the bilayer compared with a saturated sn-2 chain at the same reduced temperature. Surprisingly, addition of five more double bonds to the sn-2 chain only slightly reduces the order of the 16:0 sn-1 chain at many positions in the bilayer compared with the single double bond. Calculating oriented spectra from a spin-lattice (T1) relaxation series of powder spectra allows one to obtain the T1 relaxation times of six positions on the acyl chain simultaneously. As an example of the utility of these molecules, we demonstrate that the dependence of the spin-lattice (T1) relaxation rate as a function of orientational order for two unsaturated phospholipids differs significantly from the corresponding fully saturated analogue. Interpreting this difference using current models of acyl chain dynamics suggests that the bilayers containing either of the two unsaturated phospholipids are significantly more deformable than bilayers made from the fully saturated phospholipid.

Side-chain dynamics of two aromatic amino acids in pancreatic phospholipase A2 as studied by deuterium nuclear magnetic resonance

The flexibility of individual amino acid side chains of pancreatic phospholipase A2 in aqueous and micellar solutions was studied with deuterium nuclear magnetic resonance (2H NMR). Bovine pancreatic phospholipase A2 was selectively deuterated at the aromatic ring systems of Trp-3 and Phe-5 and porcine pancreatic phospholipase A2 at Trp-3 only. Solid-state 2H NMR spectra of the lyophilized enzymes exhibited quadrupole splittings on the order of 130 kHz, indicating almost complete immobilization of the aromatic ring systems. Exposure to a water-saturated atmosphere did not remove these steric constraints. However, side-chain mobility could be induced for the tryptophyl residue of the bovine enzyme by dissolving this enzyme in aqueous buffer or micellar solution whereas the phenyl ring always remained immobile and served as a probe for the protein's overall rotation. Typical correlation times for the tryptophyl and phenyl aromatic ring systems in aqueous solution were 7 ps and 13 ns (at 20 degrees C), respectively. The correlation time of the phenyl ring was longer than expected for the monomeric protein (approximately 6 ns), suggesting some aggregation of the protein at the high concentrations used for the NMR measurements. Addition of a micellar solution of oleoylphosphocholine had no influence on the motional freedom of the tryptophyl residue but approximately doubled the correlation time of the phenyl ring, indicating an increase of the effective volume of the tumbling particle due to lipid-protein interaction. A different behavior was observed for the Trp-3 residue of porcine phospholipase A2.(ABSTRACT TRUNCATED AT 250 WORDS)

New biophysical techniques and their application to the study of membranes

Images

Exchange rates and numbers of annular lipids for the calcium and magnesium ion dependent adenosinetriphosphatase

A spin-labeled phospholipid is used to study lipid-protein interactions in the (Ca2+,Mg2+)-ATPase of sarcoplasmic reticulum from muscle. A novel null method is used to decompose composite electron spin resonance spectra into two components, characteristic of immobilized and mobile environments. Calculations based on a random mixing model suggest that protein-protein interactions will be relatively rare in these systems and that the immobilized lipid does not represent lipid trapped between proteins but rather represents annular phospholipid at the lipid-protein interface of the adenosinetriphosphatase. The apparent decrease in the amount of immobilized lipid with increasing temperature is shown to be consistent with lipid exchange between bulk and annulus, characterized by an exchange time of 10(-7) s at 37 degrees C. A minimum number of annular phospholipid sites of 32 and 22 are calculated at 0 and 37 degrees C, respectively.

Phospholipid composition and organization of cytochrome c oxidase preparations as determined by 31P-nuclear magnetic resonance

The molecular organization as well as the composition of the phospholipids in cytochrome c oxidase preparations (bovine heart) were investigated by 31P-nuclear magnetic resonance. In the so-called 'lipid-rich' preparation the lipids were found to form a fluid bilayer around the enzyme since the 31P-NMR spectrum was characteristic of a fast, axially symmetric motion of the phosphate groups with a chemical shift anisotropy of delta sigma = -45 ppm. In contrast, the 'lipid-depleted' cytochrome c oxidase gave rise to a broader spectrum where the motion of the phospholipids was no longer axially symmetric. Nevertheless, the total width of the spectrum was still considerably narrower than observed for immobilized phospholipids in solid crystals. Both enzyme preparations were dissolved in 1% detergent solution and used for high-resolution 31P-NMR spectroscopy. Narrow lines of about 20 Hz linewidth were obtained for both types of enzyme preparations, and well-resolved resonances could be assigned to cardiolipin, phosphatidylethanolamin and phosphatidylcholine. The major differences between lipid-rich and lipid-depleted cytochrome c oxidase were the absolute amount of phospholipid associated with the protein and the relative contribution of the individual lipid classes to the 31P-NMR spectrum. For lipid-rich cytochrome c oxidase about 130 molecules phospholipid were bound per enzyme (approx. 11 cardiolipins, 54 phosphatidylethanolamines and 64 phosphatidylcholines). For lipid-depleted cytochrome c oxidase only 6-18 lipids were bound per enzyme (1 or 2 cardiolipins, 3-8 phosphatidylethanolamines and 2-8 phosphatidylcholines). In contrast to earlier suggestions that cardiolipin is the only remaining lipid in lipid-depleted cytochrome c oxidase, the 31P-NMR studies demonstrate that all three lipids remain associated with the protein.

Membrane protein conformational change dependent on the hydrophobic environment

Two conformational states of the coat protein of the filamentous bacteriophage M13 have been detected in detergent solution by using magnetic resonance techniques. When 3-fluorotyrosine is incorporated in place of the two tyrosine residues in the protein, four 19F nuclear magnetic resonance signals are observed, two for each conformer of the protein. The equilibrium between the two forms can be modulated by pH, temperature, and detergent structure. The rate of interconversion of the isomers is rapid on the minutes time scale but is slow relative to the T1 relaxation time of the fluorine resonances of approximately 50 ms. The conformational change between the conformers results in the perturbation of a basic residue in the protein such that this group has a pKa of approximately 9.5 in one state which shifts to 10.5 or more in the other conformational state. The temperature dependence of the equilibrium suggests an enthalpy difference of about 10 kcal/mol which is offset by entropy to give nearly zero free energy difference between the states at pH 8.3 in deoxycholate solution at room temperature. This suggests a substantial reorganization of the noncovalent interactions defining the two conformational states. The conformational equilibrium is strongly dependent on detergent structure and the presence of phospholipid in the detergent micelle. The results are not consistent with a strong, specific lipid binding to the protein but appear to be consistent with a more general effect of the overall micelle structure on the conformational state of the protein.

Phospholipid exchange between restricted and nonrestricted domains in sarcoplasmic reticulum vesicles

Phosphorus nuclear magnetic resonance spectra of rabbit muscle light sarcoplasmic reticulum membranes consist of two overlapping resonances, one much broader than the other. The broad resonance arises from phospholipids motionally restricted, probably by association with the Ca2+-ATPase, while the narrow resonance arises from phospholipid only slightly perturbed by the presence of the protein. (Selinsky, B.S. and Yeagle, P.L. (1984) Biochemistry 23, 2281-2288). The rate of exchange between the two phospholipid domains represented by the resonances was determined by measuring the transfer of magnetization from the broad resonance to the narrow resonance. The rate of exchange of phospholipids from the restricted domain to the nonrestricted domain was determined to be 1 s-1.

Lipid-apolipoprotein interactions in surfactant studied by reassembly

Recent evidence suggests that the structure of tubular myelin, an extracellular form of pulmonary surfactant, is dependent on the interaction of lipids with certain proteins specific for this material and with calcium ions. In order to investigate how protein and calcium may affect the surfactant complex, we studied the composition and properties of reassembly materials formed with a major surfactant apolipoprotein (35,000-38,000 molecular weight) and the principal lipids found in the natural material. We were interested in three questions: 1) Does this apolipoprotein preferentially associate with certain of the lipids in surfactant? 2) What forces are involved in the binding? 3) Does the interaction result in changes in the physical state of the lipid? We found that this apolipoprotein binds phosphatidylcholines that are in a gel phase with much greater affinity than it does phosphatidylcholines that are liquid-crystalline. However, maximum binding does not occur with the pure phosphatidylcholines but rather with mixtures of phosphatidylcholines and 15% phosphatidylglycerol. Calcium ions have no effect on the amount of apolipoprotein that is bound, but they do modify the physical state of the reassembly lipoprotein and the stoichiometry of lipid to protein. These results indicate that the binding of the apolipoprotein to the lipid does not primarily involve ionic bonds. However, apolar interactions, which are influenced by the state of the lamellar phospholipid, appear to be important. Small amounts of phosphatidylglycerol and other glycolipids, which probably disrupt the regularity of a gel-phase lamellar structure when mixed with saturated phosphatidylcholines, may provide binding sites favoring the interaction. Indirect evidence, based on thermodynamic analyses, suggests that phosphatidylcholines may be partially immobilized about the protein in the formation of the complex. This conclusions is reinforced by the preliminary findings obtained from the differential scanning calorimetry of the reassembly materials.

Mattress model of lipid-protein interactions in membranes

A thermodynamic model is proposed for describing phase diagrams of mixtures of lipid bilayers and amphiphilic proteins or polypeptides in water solution. The basic geometrical variables of the model are the thickness of the hydrophobic region of the lipid bilayer and the length of the hydrophobic region of the proteins. The model incorporates the elastic properties of the lipid bilayer and the proteins, as well as indirect and direct lipid-protein interactions expressed in terms of the geometrical variables. The concept of mismatch of the hydrophobic regions of the lipids and proteins is an important ingredient of the model. The general phase behavior is calculated using simple real solution theory. The phase behavior turns out to be quite rich and is used to discuss previous experiments on planar aggregations of proteins in phospholipid bilayers and to propose a systematic study of synthetic amphiphilic polypeptides in bilayers of different thicknesses. The model is used to interpret the influence of the lipid-protein interaction on calorimetric measurements and on local orientational order as determined by deuterium nuclear magnetic resonance.

Dynamic structure of membranes by deuterium NMR

Progress in our understanding of the dynamic structure of membrane lipids and proteins has recently been made possible by the advent of high-field "solid-state" nuclear magnetic resonance spectroscopic studies of specifically deuterium-labeled systems. Major features of lipid and protein dynamics have been deduced.