Stacking of safranine in liposomes during valinomycininduced efflux of potassium ions

Liposomes were prepared from phosphatidylcholine and cardiolipin in a KCl medium and suspended in a choline chloride medium with safranine. When efflux of K⁺ was induced by valinomycin, spectral shifts characteristic of stacking were observed. Ca²⁺ inhibited the rate of stacking in a competitive manner with a K_i of about 200 μM, while La³⁺ was about 10 times more potent. When liposomes were prepared from phospholipids with a higher ratio of cardiolipin to phosphatidyl-choline the inhibition was more potent. No effect on the stacking phenomena was seen when Ca²⁺ was added after the stacking was completed. When Ca²⁺ or an organic cation with four charges, spermine, was trapped in the intraliposomal compartment, no significant change in the rate of stacking was seen. However, the extent of stacking was decreased. It is suggested that safranine is driven by a diffusion potential to a site that is inaccessible to Ca²⁺ in the medium, presumably to the inner boundaries of the liposomal membranes.     

Properties and subcellular localization of elastase-like activities of arterial smooth muscle cells in culture

The interaction of cardiolipin with Ca2+ was assessed by measuring the cardiolipin-mediated extraction of 45Ca2+ from an aqueous to an organic (methylene chloride) phase. Cardiolipin binds Ca2+ with high affinity [Kd(apparent) = 0.70 +/- 0.17 microM (S.D.)]. Cation-cardiolipin interactions are selective. Interaction of cardiolipin with Ca2+ is insensitive to Na+, but is inhibited by divalent cations with Mn2+ greater than Zn2+ greater than Mg2+. In addition La3+ and Ruthenium red are particularly potent inhibitors of Ca2+ binding by cardiolipin. Cardiolipin-mediated extraction of Ca2+ into an aqueous phase is also inhibited by phosphatidylcholine. Inhibition of Ca2+-cardiolipin interaction by phosphatidylcholine (a phospholipid known to stabilize the bilayer conformation) may implicate inverted, non-bilayer lipid structures in the binding.     

Uptake of safranine by cardiac mitochondria. Competition with calcium ions and dependence on anions.

Some characteristics of the energized uptake of safranine by rat heart mitochondria were studied. When monitored by changes in differential absorbance (between 524 and 554 nm) of a whole suspension in safranine-containing medium, the changes seen are not linearly related to the quantity of the safranine moving. It happens coincidentally that the changes observed are nearly linearly related to the logarithm of the ratio between the accumulated safranine and its residual concentration in the medium; this explains why the changes of absorbance have been found by other authors to be linearly related to the logarithms of the ratio of internal/external concentrations of such other cations as are permeable. The uptake process appears to compete for energy with Ca2+ uptake and vice versa. Energized safranine uptake has an anion requirement, which is seen when movement of endogenous Pi has been inhibited; the small residual safranine uptake obtained when energy is provided in the presence of mersalyl may be attributable to internal Pi. However, a limited anion-independent energized uptake of safranine, in exchange for internal K+, may be elicited in the presence of nigericin. Adding ATP to the energized system in the presence of an inhibitor of Pi movement elicits an additional uptake of safranine that is oligomycin-sensitive and that probably arises on account of generation of internal Pi by hydrolysis of the entering ATP.     

Translocation and binding of adenine nucleotides by rat liver mitochondria partially depleted of phospholipids.

1. Rat liver mitochondria were partially depleted of their phospholipids using phospholipase A prepared from porcine pancreas (substrate specificity, cardiolipin greater than phosphatidylethanolamine greater than phosphatidylcholine) or from Crotalus adamanteus venom (substrate specificity, phosphatidylethanolamine = phosphatidylcholine greater than cardiolipin). 2. Removal of only about 1% of the mitochondrial phospholipid with the pancreatic enzyme leads to 50% and 25% losses in ADP and ATP translocation, respectively. Concomitant with the loss in translocation is a decline in the ability of both carbonylcyanide m-chlorophenylhydrazone and Ca2+ to stimulate ATP translocation. 3. To achieve comparable losses in ADP and ATP translocation with the venom enzyme, it is necessary to remove about 8% of the total mitochondrial phospholipid. Following such treatment, carbonylcyanide m-chlorophenylhydrazone and Ca2+ are still capable of stimulating ATP translocation. 4. Control experiments involving treatment of the mitochondria with the products of phospholipase digestion indicate that the effects observed on the translocase reflect a loss of phospholipid from the membrane. 5. Binding studies indicate that the loss in adenine nucleotide translocation following phospholipase treatment cannot be accoundted for by an altered ability to bind adenine nucleotides to atractyloside-sensitive sites. 6. The data are interpreted in terms of a mechanism of adenine nucleotide translocation involving a lipoprotein carrier system, consisting of the translocator protein and phospholipids, possibly cardiolipin and phosphatidylethanolamine.     

Reconstitution of ion transport and respiratory control in vesicles formed from reduced coenzyme Q-cytochrome c reductase and phospholipids.

Reduced coenzyme Q-cytochrome c reductase from bovine heart mitochondria (complex III) was incorporated into phospholipid vesicles by the cholate dialysis procedure. Soybean phospholipids or mixtures of purified phosphatidylcholine, phosphatidylethanolamine, and cardiolipin could be used. Oxidation of reduced coenzyme Q2 by the reconstituted vesicles with cytochrome c as oxidant showed the following energy-coupling phenomena. 1. Protons were translocated outward with a coupling ratio, H+/2e, of 1.9 +/- 0.2. Measurements with mitochondria under similar conditions showed an H+/2e ratio of 1.8. Proton translocation was not seen in the presence of uncoupling agents and was in addition to the net acidification of the medium from the over-all oxidation reaction. 2. Potassium ions were taken up by the reconstituted vesicles in the presence of valinomycin in a reaction coupled to electron transfer. The coupling ratio for K+ uptake, K+/2e, was 2.0 in the vesicles and approximately 1.5 in mitochondria. 3. The rate of oxidation of reduced coenzyme Q2 by the reconstituted vesicles was stimulated up to 10-fold by uncouplers or by valinomycin plus nigericin and K+ ions. Addition of valinomycin alone in a K+ medium caused a transient stimulation of electron transfer. The results indicate that energy coupling can be observed with isolated reduced coenzyme Q-cytochrome c reductase if the enzyme complex is properly incorporated into a phospholipid vesicle.     

Spin-label studies on the specificity of interaction of cardiolipin with beef heart cytochrome oxidase

The selectivity of interaction of various cardiolipin analogues with beef heart cytochrome oxidase in reconstituted complexes with dimyristoylphosphatidylcholine has been studied by electron spin resonance spectroscopy, using lipids spin-labeled in the acyl chains. No difference in selectivity is observed between cardiolipin and its monolyso derivative, and similarly no selectivity is observed between phosphatidylcholine and lysophosphatidylcholine. Removal of the cardiolipin charge by methylation of the phosphate groups reduces but does not eliminate selectivity relative to phosphatidylcholine. The dependence of the lipid selectivity on head group and chain composition is in the order cardiolipin approximately equal to monolysocardiolipin greater than acylcardiolipin greater than dimethylcardiolipin greater than phosphatidylcholine approximately equal to lysophosphatidylcholine, where acylcardiolipin has the spin-label chain attached at the center -OH of the head group. The degree of association of the negatively charged cardiolipin derivatives with cytochrome oxidase decreases with increasing salt concentration, to a level comparable to that for dimethylcardiolipin. At high ionic strength there is still a marked selectivity relative to phosphatidylcholine. Li+ ions are more effective in screening the interaction than are Na+ ions, and divalent ions are more effective than monovalent ions. The selectivity for cardiolipin is only slightly reduced on titrating the protein to high pH. Alkylation of the protein with N-ethylmaleimide has little effect on the titration behavior. Covalent modification of the protein by reaction with citraconic anhydride decreases the selectivity of interaction with cardiolipin. It is concluded that cardiolipin possesses an additional specificity of interaction with cytochrome oxidase other than that of purely electrostatic origin.     

Effects of ionophores and metabolic inhibitors on the mitochondrial membrane potential within isolated hepatocytes as measured with the safranine method.

A difference spectrum with a peak of absorbance at 526nm appears slowly upon addition of valinomycin or KCN in combination with oligomycin to a hepatocyte suspension in the presence of safranine. When the cells are incubated at 37 degrees C in a medium containing safranine, a slow decrease in the absorbance occurs at the wavelength pair 524-484 nm. The change in absorbance is completed within 20-30 min after additions of cells to a medium containing safranine. At this time the safranine concentration of the outer medium is considerably decreased. The safranine signal is completely reversed by valinomycin, carbonyl cyanide p-trifluoromethoxyphenyl-hydrazone or KCN in combination with oligomycin. None of these treatments have any immediate effect on cellular ATP concentrations or the 36Cl- equilibrium potential across the plasma membrane. In the presence of iodoacetate a slow reversal of the trace can be induced upon addition of KCN, but not of oligomycin alone. Rotenone, in combination with oligomycin, does not reverse the safranine signal except when both KF and iodoacetate are present, in which case a slow reversal is seen. A subsequent addition of duroquinone brings back the signal to the same level as in the presence of rotenone alone. The results indicate that the spectral response of safranine in the presence of isolated hepatocytes is a result of a slow penetration of safranine into intracellular mitochondria, where aggregation of safranine molecules occurs as a response to the mitochondrial membrane potential.     

Calcium-induced fusion of proteoliposomes and protein-free liposomes Effect of their phosphatidylethanolamine content on the structure of fused vesicles

The acidic phospholipid cardiolipin was shown to be very efficient in promoting calcium-induced fusion of proteoliposomes. The degree of fusion was dependent on the phosphatidylethanolamine content of the vesicles. Addition of CaCl₂ to proteoliposomes containing phosphatidylcholine and cardiolipin but without phosphatidylethanolamine did not induce fusion. Fusion of cytochrome oxidase vesicles, containing less than 50 mol% phosphatidylethanolamine resulted in monolamellar vesicles with a diameter of about 200 nm. The vesicles could be induced to fuse further by establishing an osmotic pressure across their membranes. When proteoliposomes containing more than 50 mol% phosphatidylethanolamine were fused, large vesicles with a diameter exceeding 1 μm were formed. They appeared in the electron microscope as a mixture of multilamellar and monolamellar vesicles. Fusion of corresponding liposomes resulted in formation of even larger structures appearing as dense multilamellar bodies and paracrystalline honeycomb-like lattices.     

Evidence suggesting the importance of fatty acids and the fatty acid moieties of sperm membrane phospholipids in the acrosome reaction of guinea pig spermatozoa

When guinea pig spermatozoa were preincubated 1 hr in Ca2+-free medium containing dilysocardiolipin (100--125 micrograms/ml) then exposed to Ca2+, the majority underwent an immediate acrosome reaction. Monolysocardiolipin was much less effective and the native cardiolipin was totally ineffective. Some fatty acids added to the medium could also render the spermatozoa capable of undergoing their acrosome reactions, arachidonic acid in methyl ester form being most potent. It is known that sperm membrane contains phospholipase A which cleaves membrane phospholipids into lysophospholipids and fatty acids. Most lysophospholipids are known to be potent acrosome reaction-promoting agents. As some forms of fatty acids are also potent acrosome reaction-promoting agents, both products of membrane phospholipid hydrolysis by phospholipase A (i.e., both fatty acids and lysophospholipids) may work synergistically to effect the conversion of stable sperm membranes to a fusion competent state capable of engaging in the acrosome reaction.     

Further studies on the cardiolipin phosphodiesterase of Escherichia coli.

The cardiolipin phosphodiesterase of Escherichia coli was further characterized. This enzyme has a pH optimum of 7.0 and is Mg2+ dependent. Mn2+ and Co2+ could replace Mg2+ but other divalent cations were inhibitory or without effect. The enzyme is not periplasmic and does not appear to be associated with membrane fractions prepared by different methods. It is recovered as a soluble protein in the cytosol fraction but could not be readily purified because of its instability. With cell-free systems, a requirement for ATP or ADP could be shown under certain defined conditions. Other nucleotides were less effective or ineffective in stimulating the phosphodiesterase. The cells displayed the highest activity during the middle to late exponential stage but no marked requirement for ATP was apparent when the phosphodiesterase was obtained from such freshly grown cells. If, however, cells were starved for several hours in saline medium, the cardiolipin phosphodiesterase level fell and a requirement for added ATP could be shown. The cardiolipin phosphodiesterase is an enzyme distinct from cardiolipin synthase. The assay conditions are quite different from each of these enzymes as are their subcellular distributions.     

Calcium dependence of human sperm fertilizing ability

The Ca2+ dependency of human-sperm fertilizing ability was investigated using a modified Tyrode's medium either containing 2.4 mM CaCl2 (CA medium) or with the CaCl2 replaced by SrCl2 and 0.1 mM EGTA added to chelate any residual Ca2+ ions (SREG-medium). Ten washed sperm populations incubated in either medium for 0, 6, and 22 hours showed the same occurrence of acrosome reactions (by fluorescent lectin labelling and triple stain). A further 3-hour incubation after washing into fresh CA medium resulted in only a slight increase in acrosome reactions in both media. Eight sperm populations preincubated overnight in CA and SREG media were coincubated for 1 hour with previously salt-stored human zonae pellucidae also in the same media. Significantly more motile spermatozoa were bound to more of the zonae in CA medium (53.9% vs. 27.6% of zonae with 13.8 vs. 4.3 sperm/bound zona). In three hamster egg penetration test (HEPT) experiments, sperm populations preincubated overnight in either CA or SREG media were coincubated with hamster oocytes prepared in the same media. Only 2.1% of oocytes (1.0 polyspermy) were penetrated in SREG medium, cf., 30.9% of oocytes (1.3 polyspermy) in CA medium. These results demonstrate that while Sr2+ ions can substitute fully for Ca2+ in the capacitation and acrosome reaction of human spermatozoa, sperm-zona, and sperm-oolemma interaction seem to involve some more Ca2+-specific process(es). Furthermore, the increased HEPT fertilizing ability of human spermatozoa using overnight preincubation in SREG medium and CA medium for the test cannot be explained on the basis of differential kinetics of capacitation or the acrosome reaction.     

The enzymatic function of tafazzin

Tafazzin is a putative enzyme that is involved in cardiolipin metabolism, it may carry mutations responsible for Barth syndrome. To identify the biochemical reaction catalyzed by tafazzin, we expressed the full-length isoform of Drosophila melanogaster tafazzin in a baculovirus-Sf9 insect cell system. Tafazzin expression induced a new enzymatic function in Sf9 cell mitochondria, namely 1-palmitoyl-2-[14C]linoleoyl-phosphatidylcholine:monolysocardiolipin linoleoyltransferase. We also found evidence for the reverse reaction, because tafazzin expression caused transfer of acyl groups from phospholipids to 1-[14C]palmitoyl-2-lyso-phosphatidylcholine. An affinity-purified tafazzin construct, tagged with the maltose-binding protein, catalyzed both forward and reverse transacylations between cardiolipin and phosphatidylcholine, but was unable to utilize CoA or acyl-CoA as substrates. Whereas tafazzin supported transacylations between various phospholipid-lysophospholipid pairs, it showed the highest rate for the phosphatidylcholine-cardiolipin transacylation. Transacylation activities were about 10-fold higher for linoleoyl groups than for oleoyl groups, and they were negligible for arachidonoyl groups. The data show that Drosophila tafazzin is a CoA-independent, acyl-specific phospholipid transacylase with substrate preference for cardiolipin and phosphatidylcholine.     

On the mechanism of the phospholipase C-mediated attenuation of cardiolipin biosynthesis in H9c2 cardiac myoblast cells

The effect of phospholipase C treatment on cardiolipin biosynthesis was investigated in intact H9c2 cardiac myoblasts. Treatment of cells with phosphatidylcholine-specific Clostridium welchii phospholipase C reduced the pool size of phosphatidylcholine compared with controls whereas the pool size of cardiolipin and phosphatidylglycerol were unaffected. Pulse labeling experiments with [1,3-³H]glycerol and pulse-chase labeling experiments with [1,3-³H]glycerol were performed in cells incubated or pre-incubated in the absence or presence of phospholipase C. In all experiments, radioactivity incorporated into cardiolipin and phosphatidylglycerol were reduced in phospholipase C-treated cells with time compared with controls indicating attenuated de novo biosynthesis of these phospholipids. Addition of 1,2-dioctanoyl-sn-glycerol, a cell permeable 1,2-diacyl-sn-glycerol analog, to cells mimicked the inhibitory effect of phospholipase C on cardiolipin and phosphatidylglycerol biosynthesis from [1,3-³H]glycerol indicating the involvement of 1,2-diacyl-sn-glycerol. The mechanism for the reduction in cardiolipin and phosphatidylglycerol biosynthesis in phospholipase C-treated cells appeared to be a decrease in the activities of phosphatidic acid:cytidine-5triphosphate cytidylyltransferase and phosphatidylglycerolphosphate synthase, mediated by elevated 1,2-diacyl-sn-glycerol levels. Upon removal of phospholipase C from the incubation medium, phosphatidylcholine biosynthesis from [methyl-³H]choline was markedly stimulated. These data suggest that de novo phosphatidylglycerol and cardiolipin biosynthesis may be regulated by 1,2-diacyl-sn-glycerol and support the notion that phosphatidylglycerol and cardiolipin biosynthesis may be coordinated with phosphatidylcholine biosynthesis in H9c2 cardiac myoblast cells.     

Cardiolipin activates cytochrome c peroxidase activity since it facilitates H2O2 access to heme

In this work, the effect of liposomes consisting of tetraoleyl cardiolipin and dioleyl phosphatidylcholine (1 : 1, mol/mol) on the rate of three more reactions of Cyt c heme with H2O2 was studied: (i) Cyt c (Fe2+) oxidation to Cyt c (Fe3+), (ii) Fe...S(Met80) bond breaking, and (iii) heme porphyrin ring decomposition. It was revealed that the rates of all those reactions increased greatly in the presence of liposomes containing cardiolipin and not of those consisting of only phosphatidylcholine, and approximately to the same extent as peroxidase activity. These data suggest that cardiolipin activates specifically Cyt c peroxidase activity not only because it promotes Fe...S(Met80) bond breaking but also facilitates H2O2 penetration to the reaction center.     

Spin label studies on the selectivity of lipid-protein interaction of cardiolipin analogues with the Na+/K+-ATPase

The selectivity of lipid-protein interaction for various spin-labelled cardiolipin analogues in Na+/K+-ATPase membranes from Squalus acanthias has been investigated by ESR spectroscopy. Cardiolipin derivatives with different numbers of acyl chains, or in which the headgroup charge has been removed by methylation of the phosphate groups, all show a pronounced selectivity relative to phosphatidylcholine. Maximally three times more of the cardiolipin analogue is associated with the protein, than is phosphatidylcholine. The selectivity pattern in the absence of salt is in the order: cardiolipin approximately monolysocardiolipin greater than or equal to acylcardiolipin greater than dimethylcardiolipin much greater than phosphatidylcholine, where acylcardiolipin has the spin label chain attached to the centre-OH group of the headgroup. The degree of association of the negatively charged cardiolipins with the protein is reduced by salt, corresponding to the lower selectivity for dimethylcardiolipin. It is concluded that the selectivity of the Na+/K+-ATPase for cardiolipin is not solely of electrostatic origin, nor is it likely to originate in the larger number of fatty acid chains relative to diacyl phospholipids.     

Phospholipids of Nocardia coeliaca

The lipids of Nocardia coeliaca were separated into at least 10 components by the use of thin-layer chromatography. Phosphatidylcholine was the most abundant phospholipid in this organism, accounting for 25 to 40% of the total phospholipids. The major fatty acid components of the phosphatidylcholine were 14-methyl-pentadecanoic acid (41%), the other C(15) and C(17) iso- and anteiso-fatty acids (29%), and palmitic acid (13.5%). The next most abundant phospholipid was phosphatidylethanolamine (25 to 30%), followed by phosphatidylinositol (11 to 14%) and cardiolipin (7 to 15%). Phosphatidylethanolamine and phosphatidylinositol were very similar to the phosphatidylcholine in fatty acid composition, whereas cardiolipin was characterized by a higher content of palmitic acid (30%). In all of the phospholipids examined, only trace amounts of monounsaturated fatty acids were present. When washed cells of N. coeliaca were incubated with methionine-methyl-(14)C for 1 to 3 hr, the radioactivity was mainly incorporated into the choline moiety of the phosphatidylcholine. In contrast, acetate-1-(14)C or glycerol-1-(14)C was incorporated much more slowly into the phosphatidylcholine than into the other phospholipids and neutral lipids. No phosphatidylcholine was detected in 10 other species of Nocardia examined.     

Passive Ca2+ permeability of phospholipid vesicles and sarcoplasmic reticulum membranes.

During the excitation of muscle the estimated rate of Ca2+ release from sarcoplasmic reticulum may increase 10(3)- to 10(4)-fold compared with relaxed muscle or isolated sarcoplasmic reticulum in vitro, implying a major change in the calcium permeability of the sarcoplasmic reticulum membrane. As a first step in the assessment of the role of various membrane constituents in the regulation of calcium fluxes, the contribution of phospholipids to the definition of calcium permeability was studied in model systems. The rate of calcium release from vesicles prepared from pure phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine, phosphatidylinositides, cardiolipin, and extracted microsomal lipids is in the range of 10(-15) to 10(18) mol of calcium/cm2/s. This rate is several orders of magnitude lower than the passive calcium outflux from isolated sarcoplasmic reticulum membranes. The permeability to Ca2+ is influenced by fatty acid composition and net charge and it is markedly increased with increasing temperature or after the addition of local anesthetics.     

Influence of sterols and phospholipids on sarcolemmal and sarcoplasmic reticular cation transporters

We have examined the influence of different sterols and phospholipids on the activities of the cardiac sarcolemmal Na+-Ca2+ exchanger and Na+,K+-ATPase and the sarcoplasmic reticular Ca2+-ATPase in reconstituted proteoliposomes. When either the solubilized Na+-Ca2+ exchanger or the Na+,K+-ATPase is reconstituted into phosphatidylcholine (PC):phosphatidylserine (30:50 by weight) vesicles, high cholesterol levels (20% by weight) are required for activity to be expressed. This sterol requirement is highly specific for cholesterol. Several cholesterol analogues with minor structural changes are unable to support Na+-Ca2+ exchange or Na+,K+-ATPase activities. When solubilized sarcolemma is reconstituted into PC:cardiolipin vesicles, however, the requirement for cholesterol is lost. Substantial activity can be obtained in the complete absence of cholesterol or in the presence of several cholesterol analogues. Thus, sterol/protein interactions can be highly dependent on the phospholipid environment. In contrast, the skeletal muscle sarcoplasmic reticular Ca2+-ATPase functions equally well in the presence or absence of cholesterol after reconstitution into either PC:phosphatidylserine or PC:cardiolipin proteoliposomes. Phospholipid requirements of the transporters were also examined. The sarcolemmal Na+-Ca2+ exchanger, Na+,K+-ATPase, and the sarcoplasmic reticular Ca2+-ATPase all function optimally in the presence of phosphatidylserine or cardiolipin after reconstitution. Thus, the sarcolemmal cation transporters have similar sterol and phospholipid requirements and may have structural similarities in their hydrophobic regions. The sarcoplasmic reticular Ca2+ pump evolved in a low cholesterol membrane and has different lipid interactions. These findings may have general applicability to other plasma membrane and endoplasmic reticular enzymes.     

ADP/ATP carrier protein from beef heart mitochondria has high amounts of tightly bound cardiolipin, as revealed by 31P nuclear magnetic resonance

An unusual binding of cardiolipin to the ADP/ATP carrier has been found, which is distinguished by the relatively large amount and by the tightness of binding. High-resolution 31P NMR studies on the detergent-solubilized ADP/ATP carrier from beef heart mitochondria revealed narrow signals from phosphatidylcholine and phosphatidylethanolamine and a broadened signal of 30-40-Hz line width, suggestive of cardiolipin. Line broadening of this magnitude is to be expected when tumbling of the whole protein-detergent micelle is the only source of phosphorus spin-spin relaxation. Thus a strong immobilization of the protein-bound cardiolipin is inferred. By sucrose density gradient centrifugation phosphatidylcholine and phosphatidylethanolamine were removed, while approximately six +/- one molecules of cardiolipin remained tightly bound in the dimeric protein molecule. The cardiolipin binding was stable against treatment with sodium dodecyl sulfate although release of the inhibitor carboxyatractyloside revealed at least partial protein denaturation. Ca2+ ions did not readily interact either with the bound cardiolipin. Complete detachment of the bound phospholipid was achieved by a short heat pulse in the presence of sodium dodecyl sulfate. Denaturation of the carrier protein by guanidinium chloride or NaClO4 also led to release of the bound phospholipid. Thus different stages of protein denaturation must be envisaged.     

Lipid-protein interactions in ADP-ATP carrier/egg phosphatidylcholine recombinants studied by spin-label ESR spectroscopy

The stoichiometry and specificity of lipid-protein interaction, as well as the lipid exchange rates at the protein interface, have been determined from the electron spin resonance spectra of spin-labeled lipids in reconstituted complexes of the mitochondrial ADP-ATP carrier with egg phosphatidylcholine. With the exception of cardiolipin and phosphatidic acid, the lipids studied are found to compete for approximately 50 sites at the intramembranous surface of the protein dimer. This number of first-shell lipid sites is unusually large for a protein of this size. The specificity for the protein is in the order stearic acid approximately phosphatidic acid approximately cardiolipin greater than phosphatidylserine greater than phosphatidylglycerol approximately phosphatidylcholine, with the maximum association constant relative to phosphatidylcholine being approximately 4. The selectivity for anionic lipids was partially screened with increasing ionic strength, but to a lesser extent for cardiolipin and phosphatidic acid than for stearic acid. Only in the case of phosphatidylserine was the selectivity reduced at high ionic strength to a level close to that for phosphatidylcholine. The off rates for lipid exchange at the protein surface were independent of lipid/protein ratio and correlated in a reciprocal fashion with the different lipid selectivities, varying from 5 x 10(6) s-1 for stearic acid at low ionic strength to 2 x 10(7) s-1 for phosphatidylcholine and phosphatidylglycerol. The off rates for cardiolipin were unusually low in comparison with the observed selectivity, and indicated the existence of a special population of sites (ca. 30% of the total) for cardiolipin, at which the exchange rate was very low.(ABSTRACT TRUNCATED AT 250 WORDS)