Proton and carbon-13 nuclear magnetic resonance studies of rhodopsin-phospholipid interactions

Proton and carbon-13 nuclear magnetic resonance (1H and 13C NMR) spectra of rhodopsin-phospholipid membrane vesicles and sonicated disk membranes are presented and discussed. The presence of rhodopsin in egg phosphatidylcholine vesicles results in homogeneous broadening of the methylene and methyl resonances. This effect is enhanced with increasing rhodopsin content and decreased by increasing temperature. The proton NMR data indicate the phospholipid molecules exchange rapidly (less than 10(-3) s) between the bulk membrane lipid and the lipid in the immediate proximity of the rhodopsin. These interactions result in a reduction in either or both the frequency and amplitude of the tilting motion of the acyl chains. The 13C NMR spectra identify the acyl chains and the glycerol backbone as the major sites of protein lipid interaction. In the disk membranes the saturated sn-1 acyl chain is significantly more strongly immobilized than the polyunsaturated sn-2 acyl chain. This suggest a membrane model in which the lipid molecules preferentially solvate the protein with the sn-1 chain, which we term an edge-on orientation. The NMR data on rhodopsin-asolectin membrane vesicles demonstrate that the lipid composition is not altered during reconstitution of the membranes from purified rhodopsin and lipids in detergent.     

Effect of phospholipid removal on the kinetics of the metarhodopsin I to metarhodopsin II reaction.

The kinetics of the metarhodopsin (meta) I → metarhodopsin II reaction have been studied by flash photolysis in two different types of preparations of bovine rhodopsin: (i) digitonin-solubilized rod outer segment (ROS) membranes with a molar ratio of phospholipid to rhodopsin of approximately 90, and (ii) digitonin-solubilized phospholipid-free rhodopsin with a molar ratio of phospholipid to rhodopsin of less than 0.2. At 20 °C the kinetics in both preparations are multiexponential, but four terms are required to fit the data with the solubilized membranes, whereas only two are required with the phospholipid-free preparation. Thus, phospholipid removal simplifies the kinetics of the meta I → meta II reaction, but the resulting preparation still does not show first-order kinetics. The ratio of the time constants of these two components with detergent-solubilized phospholipid-free rhodopsin was nearly equal to the values found with ROS particles, rhodopsin-phospholipid recombinants and intact rabbit eyes. This suggests a common origin for these two components in all these preparations and appears to exclude heterogeneity in bound phospholipid as the basis of these two-component kinetics.     

Reassembly in vitro of lung surfactant lipoprotein

This report describes a successful attempt to reassemble, $\text{in vitro}$, two fractions obtained from bovine lung surfactant lipoprotein. An apoprotein isolated by gel filtration in the presence of sodium deoxycholate was recombined with lipid extracts of the surfactant, in a highly alkaline buffer (pH 10) containing 10 mM sodium deoxycholate. Sonication, dilution 1 to 10, dialysis, and washing by means of centrifugation were used to produce a lipid-protein complex. Centrifugation in a continuous sucrose density gradient revealed that this material had a density of 1.081 gm/ml and a phospholipid/protein ratio respectively almost the same as those of the original lipoprotein.     

Spin-label electron paramagnetic resonance and differential scanning calorimetry studies of the interaction between mitochondrial cytochrome c oxidase and adenosine triphosphate synthase complex.

The interaction between cytochrome c oxidase complex and adenosine triphosphate synthase (F1F0) complex in the purified, dispersed state and embedded in phospholipid vesicles was studied by differential scanning calorimetry and by spin-label electron paramagnetic resonance. The detergent-dispersed cytochrome oxidase and F1F0 complexes undergo endothermic thermodenaturation. However, when these complexes are embedded in phospholipid vesicles, they undergo exothermic thermodenaturation. The energy released is believed to result from the collapse of a strained interaction between unsaturated fatty acyl groups of phospholipids and an exposed area of the complex formed by the removal of interacting proteins. The exothermic enthalpy change of thermodenaturation of a protein-phospholipid exothermic enthalpy change of thermodenaturation of a protein-phospholipid vesicle containing both cytochrome oxidase complex and F1F0 was smaller than that of a mixture of protein-phospholipid vesicles formed from each individual electron transfer complex. This suggests specific interaction between cytochrome oxidase complex and F1F0 in the membrane. Further evidence for interaction between these two complexes is provided by saturation transfer EPR studies in which the rotational correlation time of spin-labeled cytochrome oxidase increases significantly when the complex is mixed with F1F0 prior to being embedded in phospholipid vesicles. From these results, it is concluded that at least a part of cytochrome oxidase and a part of F1F0 form a supermacromolecular complex in the inner mitochondrial membrane. No such supermacromolecular complex is detected between F1F0 and ubiquinol--cytochrome c reductase.     

Absorption of surfactants by membranes: erythrocytes versus synthetic vesicles.

Three surfactants (chlorpromazine hydrochloride, thioridazine hydrochloride, and sodium deoxycholate) are found to absorb just as strongly into the protein-containing membranes of erythrocytes as into the phospholipid bilayers of synthetic vesicles. In the concentration region where hemolysis occurs and the Langmuir adsorption isotherm is no longer valid, one may use a phase partition model in which the erythrocyte membrane is one of the phases. The partition coefficients, expressed as the ratio of mole fraction surfactant in the membrane lipid phase to concentration of surfactant in the aqueous phase, have been calculated at the point of saturation in the erythrocyte membrane. These values are Ky = 430 M-1 (chlorpromazine, pH 5.9), 550 M-1 (deoxycholate, pH 7.6), and 640 M-1 (thioridazine, pH 5.9), in isotonic buffer at 27 degrees C. Corresponding values for synthetic vesicles made from dimyristoylphosphatidylcholine are Kx = 230 M-1 (chlorpromazine, 0.12 M buffer/KCl pH 5.9), 440 M-1 (deoxycholate, 0.20 M buffer/NaCl pH 8.0) and 510 M-1 (thioridazine, 0.12 M buffer/KCl pH 5.9), at 27 degrees C. It appears that the surfactants become an integral part of the bilayer in both vesicles and natural membranes and that the absorption is not of a peripheral nature. There is no evidence that the presence of proteins in the natural membrane inhibits the absorption of these surfactants in any way.     

A new method for making phospholipid vesicles and the partial reconstitution of the (Na+, K+-activated ATPase

It has been found that vesicles of phospholipid (96% (w/w) phosphatidylcholine; 4% (w/W) phosphatidylserine) can be formed by dialysis of a solution of the phospholipid in the detergent, sodium deoxycholate. Depending upon the composition of the dialysis medium, small closed vesicles apparently bounded by one or two membranes or large multi-walled structures are produced. The former are predomiant if only univalent ions are present in the dialysis buffer. As the Mg²⁺ concentration is raised above about 0.1 mM multiwalled structures are found.The (Na⁺,K⁺)-ATPase (EC 3.6.1.3) from cattle brain microsomes has been solubilized with deoxycholate. Dialysis of this material after the addition of the above phospholipid mixture in detergent also produces membrane-bound vesicles. Sucrose density gradient centrifugation has been used to demonstrate that the phospholipid, (Na⁺,K⁺)-ATPase and protein reaggregate together only if the phospholipid and solubilized protein are mixed before dialysis. This method of forming artificial membranes may be a useful way of studying transport proteins in isolation as the vesicles appear to be small and closed.     

A molecular model of cooperative binding of Ca2+ with rhodopsin molecules in photoreceptor membranes]

Kinetics of calcium binding by photoreceptor membranes of cattle retina in concentration Ca2+ 0.5 and 1.0.10(-5) M in 5 mM tris-HCl buffer, pH 7.4 at 37 degrees C has been studied. Such kinetics is of oscillating nature. Analysis of calcium binding process curves by photoreceptor membranes allow to conclude, that crystalline areas of rhodopsin (receptor domains) can be formed in the structure of photoreceptor membranes. Conformation states and structure of rhodopsin molecules Ca-binding sites in receptor domains depend on the presence of Ca2+ in the medium. The structure of rhodopsin molecules Ca-binding sites in receptor domain formed in the presence of Ca2+ in the medium was proposed. According to the Hodgkin and Huxley conception concerning the properties of Na(+)- and K(+)-channels, the receptor domain with such a structure of rhodopsin molecules Ca-binding sites can represent the conjugate system of Na(+)- and K(+)-channels. Molecular mechanisms of photoreceptor and nerve cells excitation was also proposed.     

Spin-label electron paramagnetic resonance and differential scanning calorimetry studies of the interaction between mitochondrial succinate-ubiquinone and ubiquinol-cytochrome c reductases

The interaction between succinate-ubiquinone and ubiquinol-cytochrome c reductases in the purified, dispersed state and in embedded phospholipid vesicles was studied by differential scanning calorimetry and by electron paramagnetic resonance (EPR). When the purified, detergent-dispersed succinate-ubiquinone reductase, ubiquinol-cytochrome c reductase, and cytochrome c oxidase undergo thermodenaturation, they show an endothermic transition. However, when these isolated electron-transfer complexes are embedded in phospholipid vesicles, they undergo exothermodenaturation. The energy released could result from the collapse of the strained interaction between unsaturated fatty acyl groups of phospholipids and an exposed area of the complex formed by removal of interacting proteins. The exothermic enthalpy change of thermodenaturation of a protein-phospholipid vesicle containing both succinate-ubiquinone and ubiquinol-cytochrome c reductases was smaller than that of a mixture of protein-phospholipid vesicles formed from the individual electron-transfer complexes. This suggests specific interaction between succinate-ubiquinone reductase and ubiquinol-cytochrome c reductase in the membrane. This idea is supported by saturation transfer EPR studies showing that the rotational correlation time of spin-labeled ubiquinol-cytochrome c reductase is increased when mixed with succinate-ubiquinone reductase prior to embedding in phospholipid vesicles. These results indicate that succinate-ubiquinone reductase and ubiquinol-cytochrome c reductase are indeed present in the membrane as a supermacromolecular complex. No such supermacromolecular complex is detected between NADH-ubiquinone and ubiquinol-cytochrome c reductases or between succinate-ubiquinone and NADH-uniquinone reductases.     

The molecular weight of rhodopsin and the nature of the rhodopsin-digitonin complex

The sedimentation behavior of aqueous solutions of digitonin and of cattle rhodopsin in digitonin has been examined in the ultracentrifuge. In confirmation of earlier work, digitonin was found to sediment as a micelle (D-1) with an s₂₀ of about 6.35 Svedberg units, and containing at least 60 molecules. The rhodopsin solutions sediment as a stoichiometric complex of rhodopsin with digitonin (RD-1) with an s₂₀ of about 9.77 Svedberg units. The s₂₀ of the RD-1 micelle is constant between pH 6.3 and 9.6, and in the presence of excess digitonin. RD-1 travels as a single boundary also in the electrophoresis apparatus at pH 8.5, and on filter paper at pH 8.0. The molecular weight of the RD-1 micelle lies between 260,000 and 290,000. Of this, only about 40,000 gm. are due to rhodopsin; the rest is digitonin (180 to 200 moles). Comparison of the relative concentrations of RD-1 and retinene in solutions of rhodopsin-digitonin shows that RD-1 contains only one retinene equivalent. It can therefore contain only one molecule of rhodopsin with a molecular weight of about 40,000. Cattle rhodopsin therefore contains only one chromophore consisting of a single molecule of retinene. It is likely that frog rhodopsin has a similar molecular weight and also contains only one chromophore per molecule. The molar extinction coefficient of rhodopsin is therefore identical with the extinction coefficient per mole of retinene (40,600 cm.² per mole) and the E(1 per cent, 1 cm., 500 mµ) has a value of about 10. Rhodopsin constitutes about 14 per cent of the dry weight, and 3.7 per cent of the wet weight of cattle outer limbs. This corresponds to about 4.2 x 10⁶ molecules of rhodopsin per outer limb. The rhodopsin content of frog outer limbs is considerably higher: about 35 per cent of the dry weight, and 10 per cent of the wet weight, corresponding to about 2.1 x 10⁹ molecules per outer limb. Thus the frog outer limb contains about five hundred times as much rhodopsin as the cattle outer limb. But the relative volumes of these structures are such that the ratio of concentrations is only about 2.5 to 1 on a weight basis. Rhodopsin accounts for at least one-fifth of the total protein of the cattle outer limb; for the frog, this value must be higher. The extinction (K₅₀₀) along its axis is about 0.037 cm.² for the cattle outer limb, and about 0.50 cm.² for the frog outer limb.     

Further characterization of the lipid-depleted bovine rhodopsin obtained by cholate-ammonium sulfate fractionation

The rhodopsin preparation obtained by the method of ammonium sulfate fractionation contained 3–6 mol phospholipid and about 18 mol cholate per mol rhodopsin. The purified rhodopsin had 74% helical structure and showed a visible CD spectrum different from that of rhodopsin in the membrane. The rhodopsin was stable below but denatured gradually above 20°C. The lifetime of metarhodopsin I was long in this preparation. Regeneration capacity was low and only 30% of the original rhodopsin was regenerable by addition of 11-cis-retinal after bleaching.50 mol of phosphatidylcholine were maximally bound to 1 mol rhodopsin when the purified rhodopsin was mixed with phosphatidylcholine in 0.5% cholate. The rhodopsin recombined with lipid had properties similar to those of the original rhodopsin in the membrane. Exchange of cholate for other detergents was easily performed by dialysis. The rhodopsin preparation in which cholate was exchanged for digitonin gave almost the same CD, thermal stability and regenerability as those of a native rhodopsin in the membrane but metarhodopsin I still retained its long lifetime.     

Protein-lipid interactions at membrane surfaces: a deuterium and phosphorus nuclear magnetic resonance study of the interaction between bovine rhodopsin and the bilayer head groups of dimyristoylphosphatidylcholine

Rhodopsin, isolated from bovine retinal rod outer segment disk membranes, has been reconstituted into bilayers of 1,2-dimyristoyl-sn-glycero-3-phosphocholine which was deuterated in the terminal methyl groups of the choline polar head group. By use of a mixed detergent system of cholate and octyl glucoside to solubilize the phospholipid and rhodopsin, 15 membrane complexes of predetermined phospholipid to rhodopsin mole ratios of between 350:1 and 65:1 have been produced by exhaustive dialysis and studied by a variety of techniques. Electron micrographs of replicas from freeze-fractured membrane complexes showed that the majority of the lipid, for all rhodopsin:phospholipid ratios, was contained in large bilayer vesicles with diameters in excess of 400 nm. Complexes produced with rhodopsin from frozen retina produced an absorption maximum at 478 nm after photobleaching whereas rhodopsin from fresh retina could be bleached more completely to an absorption maximum at 380 nm. Deuterium nuclear magnetic resonance (NMR) spectra from the lipid head groups of bilayers above the gel to liquid-crystalline phase transition temperature were shown to be sensitive in a systematic way to the presence of rhodopsin which could be bleached to 380 nm. The measured quadrupole splittings, taken as the separation of the turning points of the recorded NMR spectra, decreased from a value of 1.28 kHz for protein-free bilayers to approximately 0.40 kHz for bilayers containing 65 molecules of phospholipid for each rhodopsin at 32 degrees C.(ABSTRACT TRUNCATED AT 250 WORDS)     

Solubilization of low-density lipoprotein with sodium deoxycholate and recombination of apoprotein B with dimyristoylphosphatidylcholine

Apoprotein B (apoB) of human plasma low-density lipoprotein (LDL) (d 1.025-1.050 g/mL) has been solubilized with solid sodium deoxycholate (NaDC) above its critical micellar concentration. ApoB is isolated by gel-filtration chromatography as a mixed micellar complex of protein and detergent in high yield in a lipid-free form. A soluble apoB-dimyristoylphosphatidylcholine (DMPC) complex has been prepared by incubation of aqueous solutions of apoB-NaDC and DMPC-NaDC (2/1 w/w) at room temperature with detergent removal by extensive dialysis. A combination of gel chromatographic and density gradient fractionation of DMPC-apoB incubation mixtures demonstrates that a reasonably well-defined complex of DMPC and apoB is formed with a 4:1 w/w lipid:protein ratio. Negative-stain electron microscopy shows these particles to be single-bilayer phospholipid vesicles with a diameter of 210 +/- 20 A into which the apoB is incorporated. Circular dichroic spectra of NaDC-solubilized apoB show apoB to have similar conformation to that seen in the native LDL particle. However, apoB that has been complexed with DMPC exhibits more alpha-helix. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis shows a single band (apparent Mr 366000) for apoB after solubilization, purification, and interaction with phospholipid. The behavior of apoB during its reassociation with phospholipid and the structural features of the DMPC-apoB particle are similar to those observed in the interaction of solubilized membrane proteins with lipid rather than that of other apo-lipoproteins.     

Highly sequential binding of protein kinase C and related proteins to membranes

Protein kinase C belongs to a class of proteins that displays simultaneous interaction with calcium and phospholipids. Other members of this class include two proteins (Mr 64K and 32K) isolated from bovine brain. The association of these proteins with membranes exhibited highly unusual properties that were not consistent with a simple equilibrium. Titration of protein-phospholipid binding as a function of calcium showed an apparently normal curve with a low degree of cooperativity. The binding was rapid and quickly adjusted to changes in the calcium concentration. Calcium was readily exchanged from the protein-phospholipid complex. However, at each calcium concentration, membrane-bound protein was not in rapid equilibrium with free protein in solution; the half-time for dissociation exceeded 24 h. Titration of phospholipid vesicles with proteins showed different saturation levels of bound protein at different calcium concentrations. The amount of protein bound was almost entirely determined by the concentration of calcium and was virtually unaffected by the free protein concentration. These properties suggested that protein-phospholipid binding involved a sequence of steps that were each irreversible upon completion. These binding properties were consistent with high-affinity interaction between protein and phospholipid, high cooperativity with respect to calcium (N greater than or equal to 10), clustering of acidic phospholipids, and negative cooperativity with respect to protein density on the membrane. A major apparent problem with the complete titration of PKC-membrane interaction was a requirement for calcium in excess of intracellular levels. However, a highly sequential binding process showed that a number of protein-binding sites on the membrane would be saturated with calcium at physiological levels.(ABSTRACT TRUNCATED AT 250 WORDS)     

Magnetic-circular-dichroism studies of haem a and its derivatives.

1. The Thy-1 membrane glycoproteins from rat thymus and brain bound deoxycholate to 24% of their own weight as measured by equilibrium dialysis. The binding occurred co-operatively at the critical micelle concentration of deoxycholate, suggesting that the glycoproteins bind to a micelle, and not to the detergent monomer. 2. From sedimentation-equilibrium and deoxycholate-binding data the molecular weights of the glycoprotein monomers were calculated to be 18700 and 17500 for thymus and brain Thy-1 glycoprotein monomers were calculated to be 18700 and 17500 for thymus and brain Thy-1 glycoproteins respectively. The molecular weight of the polypeptide part of the glycoprotein is thus 12500. 3. In the absence of deoxycholate, brain or thymus Thy-1 glycoprotein formed large homogeneous complexes of mol. wt. 270000 or 300000 respectively. The sedimentation coefficient of these was 12.8 S. The complex was only partially dissociated by 4M-guanidinium chloride. 4. After cleavage of brain or thymus Thy-1 glycoprotein with CNBr, two peptides were clearly identified. They were linked by disulphide bonds and both contained carbohydrate. This cleavage suggests there is only one methionine residue per molecule, which is consistent with the above molecular weights and the known amino acid composition.     

Biochemical aspects of the visual process. XXXVIII. Effects of lateral aggregation on rhodopsin in phospholipase C-treated photoreceptor membranes

Treatment of bovine rod outer segments with phospholipase C leads to largely lipid-depleted membranous structures. Under these conditions rhodopsin remains spectrally intact, but its thermal stability and regeneration capacity are decreased, whereas upon illumination the metarhodopsin I to II transition is blocked. These observations can be explained on the basis of the previously demonstrated lateral aggregation of rhodopsin molecules which, on the one hand leads to a (partial) shielding of these molecules and, on the other hand, might impose constraints on the flexibility of the molecule to undergo light-induced conformational changes.Upon reconstitution of these lipid-depleted preparations with amphipathic lipids by means of a detergent dialysis procedure, the aggregates are apparently rearranged to lipid bilayer structures with complete recovery of the original rhodopsin properties. Under our conditions the nature of the polar head groups and the fatty acids is not critical in this respect. Simple addition of amphipathic lipids, without the use of detergent, restores the rhodopsin properties only in the case of rod outer segment lipids and of didecanoylphosphatidylcholine, and even then only occasionally.These results are discussed in the light of the strong analogy in properties between phospholipase C-treated rod outer segment membranes and lipid- and detergent-free rhodopsin obtained by affinity chromatography. It is concluded that rhodopsin must be in a freely dispersed state in order to function properly. Apparently, a non-specific lipid bilayer fulfills this condition for the regeneration capacity, whereas normal photolytic behaviour requires, in addition, a minimal membrane fluidity according to the observations of other investigators. Presumably, the uniquely high phospholipid unsaturation of rod outer segment membranes is important for another, as yet unassessed, function of rhodopsin or the photoreceptor membrane.     

In vitro reassembly of the membranous vesicle from Escherichia coli outer membrane components: Role of individual compnents and magnesium ions in reassembly

A method was developed for the reassembly of membranous vesicle from the sodium dcoxycholate-dissociated outer membrane components of Escherichia coli. The removal of the detergent by dialysis and the presence of Mg²⁺ were essential for the reassembly.Membrane protein alone did not form any membranous structure. Closed membranous vesicles similar to the native outer membrane were reassembled only when protein was mixed with both lipopolysaccharide and phospholipid in deoxycholate solution and subsequently dialyzed. The membrane showed a distinct trilaminar structure with a center-to-center distance between two dark lines of 53 Å, which is a characteristic of the native outer membrane. This characteristic trilaminar structure was shown to be due to the presence of lipopolysaccharide. Phospholipd was required for the vesicularization of membrane. Lipopolysaccharide and/or phospholipid formed a membranous structure in the absence of protein, while the morphology of their negatively stained sample was quite different from that of the native outer membrane unless the outer membrane protein was added to the reassembly mixture.The protein from the cytoplasmic membrane was unable to reform membranous vesicle with lipopolysaccharide and phospholipid, indicating that the reassembly system discriminated outer membrane proteins from cytoplasmic membrane proteins.     

Disassembly and characterization of the nuclear pore complex-lamina fraction from bovine liver nuclei

The nuclear pore complex-lamina (PCL), composed of nuclear pore structures attached to fibrous lamina, was isolated from bovine liver nuclei. We found that the highly aggregated PCL was disrupted and 75% of the constituent polypeptides could be solubilized by extraction for 1 h with 2% deoxycholate (DOC) and 3% 2-mercaptoethanol. While some differential solubilization was observed at lower detergent concentrations, all PCL proteins were solubilized equally at 2% DOC. The reducing agent was necessary to achieve maximum dispersal of the PCL and to prevent aggregation of the solubilized proteins. No tightly bound phospholipid or Triton X-100 could be detected in these preparations. Rapid removal of DOC, by dialysis or gel filtration, resulted in aggregation and precipitation of the PCL proteins, but the detergent could be removed by centrifugation through sucrose gradients. The sedimentation profiles indicated that the three major polypeptides, lamins A, B, and C, each sedimented as a single peak with a shoulder of more rapidly sedimenting material, possibly higher oligomeric forms. The sedimentation coefficient of lamins B and C, in the presence and absence of detergent, was 4.5 S. In the presence of DOC, lamin A had a sedimentation coefficient of 5.6 S, but this value was decreased to 4.1 S, when DOC was omitted from the gradient. These studies suggested that lamins B and C do not interact with or bind DOC, while lamin A may bind appreciable amounts of the detergent. The Stokes radii of lamins A, B, and C were found by gel filtration to be 75, 75, and 70 A, respectively. The molecular weights and frictional ratios estimated from the sedimentation and gel filtration data indicated that the lamins are dimeric, rod-shaped molecules.     

Detergent-solubilized bovine cytochrome c oxidase: dimerization depends on the amphiphilic environment

The extent to which bovine cytochrome c oxidase (COX) dimerizes in nondenaturing detergent environments was assessed by sedimentation velocity and equilibrium. In contrast to generally accepted opinion, the COX dimer is difficult to maintain and is the major oligomeric form only when COX is solubilized with a low concentration of dodecylmaltoside, i.e., approximately 1 mg/mg protein. The dimer form is intrinsically unstable and dissociates into monomers with increased detergent concentration, i.e., >5 mg/mg protein. The structure of the solubilizing detergent, however, greatly alters detergent effectiveness by inducing either monomerization or aggregation. Triton X-100 is most effective at solubilizing COX, but it destabilizes COX dimers, even at low concentration. Undecylmaltoside, decylmaltoside, and octaethyleneglycolmonododecyl ether (C(12)E(8)) are less effective at solubilizing COX. Each prevents COX aggregation at high detergent concentration, but also destabilizes the COX dimer. Other detergents, e.g., Tween 20, sodium cholate, sodium deoxycholate, CHAPS, or CHAPSO, are completely ineffective COX solubilizers and do not prevent aggregation even at 10-40 mg/mL. The transition from dimers to monomers depends on many factors other than detergent structure and concentration, e.g., protein concentration, phospholipid content and pH. We conclude that the intrinsic dimeric structure of COX can be maintained only after solubilization with low concentrations of dodecylmaltoside at near neutral pH, and even then precautions must be taken to prevent its dissociation into monomers.     

Structural study of rhodopsin in detergent micelles by small-angle neutron scattering

Monodisperse solutions of bovine rhodopsin monomers, devoid of lipid, associated with a linear polyoxyethylene alcohol detergent have been prepared. The composition and homogeneity of these complexes have been determined by hydrodynamic characterisation. Each rhodopsin molecule is associated with about 110 monomers of the detergent. These rhodopsin-detergent complexes have been studied by small-angle neutron scattering. Partial or total deuteration of the detergent, as well as variation of the ²H₂O/H₂O ratio in the solvent, were used to eliminate the detergent—solvent contrast at various protein—solvent contrasts. The size and shape of the detergent micelle and of the rhodopsin-detergent complexes were shown to be independent of solvent or detergent deuteration. Mixture of selectively deuterated detergent molecules allowed us to obtain an homogeneous scattering density for the detergent part of the micelles and therefore to eliminate totally its contribution to the scattering when it is contrast matched. Neutron scattering from rhodopsin alone was then measured even in highly deuterated solvents, with low incoherent background, as for a water-soluble protein. Supplementary neutron scattering measurements on rhodopsin-dodecyl dimethylamine oxide micelles confirmed essentially the results reported by Yeager (1975). Analysis of the neutron scattering data indicates that most of the hydrophobic residues of rhodopsin form a compact region which has zero hydration, this probably being the part which is embedded in the disc membrane, and that the unhydrated rhodopsin molecule is asymmetrically arranged with respect to the membrane. Comparison with the results of a small-angle X-ray scattering study (Sardet et al., 1976) implies that the peripheral regions on both sides of the membrane are highly hydrated. Several schematic models are discussed.     

Incorporation of membrane proteins into large single bilayer vesicles. Application to rhodopsin

A general procedure to incorporate membrane proteins in a native state into large single bilayer vesicles is described. The results obtained with rhodopsin from vertebrate and invertebrate retinas are presented. The technique involves: (a) the direct transfer of rhodopsin-lipid complexes from native membranes into ether or pentane, and (b) the sonication of the complex in apolar solvent with aqueous buffer followed by solvent evaporation under reduced pressure. The spectral properties of rhodopsin in the large vesicles are similar to those of rhodopsin in photoreceptors; furthermore, bleached bovine rhodopsin is chemically regenerable with 9-cis retinal. These results establish the presence of photochemically functional rhodopsin in the large vesicles. Freeze-fracture replicas of the vesicles reveal that both internal and external leaflets contain numerous particles approximately 80 A in diameter, indicating that rhodopsin is symmetrically distributed within the bilayer. More than 75% of the membrane area is incorporated into vesicles larger than 0.5 micron and approximately 40% into vesicles larger than 1 micron.