Interaction of microtubule proteins with phospholipid vesicles

We have examined the interaction of unilamellar dimyristoyl phosphatidylcholine liposomes with the high-speed supernate of brain homogenate and with tubulin purified through one or two cycles of microtubule assembly-disassembly. Tubulin and certian high molecular weight proteins are selectively adsorbed from these mixtures onto liposomes. The composition of adsorbed proteins is similar to that obtained during corresponding cycles of microtubule assembly, suggesting the equivalency of these processes. Adsorption induces stacking and/or fusion of liposomes into multilamellar structures indicating strong protein-lipid interaction. In addition, liposome-adsorbed tubulin forms extensive intermolecular disulfide bridges that are inert to reducing agents in the aqueous medium. The observations form a basis for further study of the distribution, function, and properties of membrane-bound tubulin.     

Interactions of divalent cations or basic proteins with phosphatidylethanolamine vesicles

Small unilamellar vesicles have been prepared from phosphatidylethanolamine by sonication of the lipid in aqueous buffers of low ionic strength and high pH. These vesicles and their interactions with various di- and trivalent cations have been characterized using freeze-fracture electron microscopy. Phosphatidylethanolamine from 4 sources was examined: Hens' yolk phosphatidylethanolamine, human grey matter phosphatidylethanolamine, Escherichia coli phosphatidylethanolamine and dimyristoyl phosphatidylethanolamine. The phosphatidylethanolamine from natural sources formed spherical, uniform 20–40 nm vesicles while dimyristoyl phosphatidylethanolamine formed larger, 70 × 25 nm, disc-shaped vesicles when sonicated above the phase transition temperature. Fusion of the unilamellar egg phosphatidylethanolamine, E. coli phosphatidylethanolamine and human grey matter phosphatidylethanolamine vesicles was induced by dialysis against buffers containing 2.0 nM Ca⁺ or 3.0 mM Mg²⁺. The fusion of the vesicles resulted in the precipitation of the lipid and the formation of multilamellar and, in some cases, hexagonal II structures. Dimyristoyl phosphatidylethanolamine vesicles were precipitated at 55°C by 1.0 mM Ca⁺ or 2.0 mM Mg²⁺. Treatment of the calcium- and magnesium-precipitated vesicles of hen's egg yolk phosphatidylethanolamine, E. coli phosphatidylethanolamine, human grey matter phosphatidylethanolamine and dimyristoyl phosphatidylethanolamine with EDTA resulted in resuspension of the lipid. The specific size and shape of the vesicles formed in this manner depends on the type of phosphatidylethanolamine and ion involved. Dialysis of the Ca⁺- and Mg²⁺-precipitated egg phosphatidylethanolamine vesicles against buffer containing no Ca⁺, Mg²⁺ or EDTA also resulted in dissociation of the precipitate and formation again of a new vesicle population. This evidence indicates that the Ca⁺ and Mg²⁺ are not strongly bound to the phosphatidylethanolamine.Egg phosphatidylethanolamine vesicles would fuse in the presence of many di- and trivalent ions. Egg phosphatidylethanolamine vesicles were precipitated by beryllium, aluminum, chromium, manganese, cobalt, nickel, copper, zinc, strontium, cadmium, barium, lanthanium, mercury and lead. The amount of ion required to precipitate the vesicles and the type of structure resulting from the fusion of the vesicles was found to be unique for each ion.Small unilamellar vesicles prepared from egg phosphatidylethanolamine were reacted with several basic proteins (cytochrome c, basic protein from human myelin, protamine, poly-l-lysine and cationically-modified ferritin). The basic proteins also initiated the fusion of egg phosphatidylethanolamine vesicles but these proteins did not fuse egg phosphatidylcholine vesicles nor did normal ferritin initiate fusion. Human myelin basic protein initiated the fusion of dimyristoyl phosphatidylethanolamine vesicles above and below the phase transition of this lipid.     

Interaction of carrier ionophores with phospholipid vesicles

The interactions of carrier ionophores, nonactin, A23187, and lasalocid A with liposomes formed from the synthetic lipids dimyristoylphosphatidylcholine and dipalmitoylphosphatidylcholine are investigated by differential scanning calorimetry and 1H and 31P nuclear magnetic resonance techniques. The results indicate that the mode of interaction of these ionophores is dependent on the fluidity of the bilayer and on the chemical nature of these ionophores. The 31P NMR studies are suggestive of the formation of small particles that are probably intervesicular lipid-ionophore aggregates in multilamellar vesicles when they are incorporated with these ionophores at high concentrations. The results are interpreted on the basis of the chemical structure and conformations of the ionophores in membrane mimetic media. The 1H NMR line-width measurements indicate that the aromatic rings containing the carboxyl groups of lasalocid A and A23187 are located near the membrane interface while the rest of the molecule is buried in the membrane interior.     

Interaction of Penicillium notatum phospholipase B with divalent cations

The interaction between Penicillium notatum phospholipase B and divalent cations such as Ca2+ and Mg2+ was studied. When the purified enzyme, present at concentrations of submicrogram to microgram per ml, was incubated with submillimolar to millimolar concentrations of CaCl2 or MgCl2, the enzymatic activity was remarkably decreased (to no more than 30% of original activity, when the enzyme was incubated with 2 mM CaCl2 for 15 min). The inhibitory effect of divalent cations was reversible, since dialysis against a metal chelator, such as EDTA or EGTA, substantially restored the enzymatic activity. Atomic absorption analysis showed the purified enzyme molecule to be present in a complex with Ca2+ at a ratio approaching 1:1, and this Ca2+ binding was shown to be extremely tight, since repeated dialyses of the enzyme molecules against EDTA or EGTA could remove the divalent cations only in a gradual manner. During this process, the enzyme activity increased also gradually. The remnant fraction of tightly bound Ca2+ was released from the enzyme molecule after the denaturation of the enzyme by treatment with guanidine hydrochloride, and the apoenzyme recovered its substantial activity after removal of the denaturing agent by dialysis. On the other hand, the content of Mg2+ in the purified enzyme molecule was lower than that of Ca2+, and the association of Mg2+ with the enzyme was much weaker in comparison to that of Ca2+. Atomic absorption analysis of the enzyme exposed to exogenous Ca2+ showed a fast removal, by dialysis, of unbound and weakly bound divalent cation, followed by a gradual removal of endogenous Ca2+ and a concomitant increase of enzymatic activity, which are similar to data obtained for the purified enzyme. Results shown in this report suggest some regulatory roles of divalent cations, especially of Ca2+, in the enzymatic function of P. notatum phospholipse B.     

Phase changes of cardiolipin vesicles mediated by divalent cations

Small unilamellar vesicles were prepared from cardiolipin and produced the hexagonal II phase when dialyzed against CaCl₂ or MgCl₂. Upon removal of the cation by dialysis against EDTA large unilamellar vesicles were formed. The events of the transition from the lamellar to hexagonal phase and back to the lamellar phase are described.     

Interaction of divalent cations with beta-galactosidase (Escherichia coli).

Although the addition of various divalent metals to beta-galactosidase resulted in apparent activation, only Mg2+ and Mn2+ actually did activate. The apparent activation by the other divalent metals was shown to be due to Mg2+ impurities. Calcium did not activate, but experiments suggested that it did bind. Other divalent metals which were studied failed to bind. The dissociation constants for Mg2+ and Mn2+ were 2.8 X 10(-7) and 1.1 X 10(-8) M, respectively, and in each case one ion bound per monomer. These constants corresponded very closely to apparent values which were obtained from activation studies. The apparent binding constant for Ca2+, obtained from competition studies, was 1.5 X 10(-5) M. Data were obtained which showed that Mg2+, Mn2+, and Ca2+ all compete for binding at a single site. Of interest and of possible molecular biological importance was the observation that, while Mg2+ bound noncooperatively (n = 1.0), Mn2+ did so in a highly cooperative manner (n = 3.4). The binding of Mn2+ (as compared to Mg2+) resulted in a twofold drop in the Vmax for the hydrolysis and transgalactosylis reactions of lactose but had little effect on the Vmax of hydrolysis of allolactose, p-nitrophenyl beta-D-galactopyranoside (PNPG), or o-nitrophenyl beta-D-galactopyranoside (ONPG); Km values were not effected differently for any of the substrates by Mn2+ as compared to Mg2+. When very low levels of divalent metal ions were present (0.01 M EDTA added) or when Ca2+ was bound with lactose as the substrate, a greater decrease was observed in the rate of the transgalactosylic reaction than in the rate of the hydrolytic reaction, and the Km values for lactose and ONPG were increased. Of the three divalent metal ions which bound to beta-galactosidase, only Mn2+ had significant stabilizing effects toward denaturing urea and heat conditions.     

Phase changes of cardiolipin vesicles mediated by divalent cations.

Small unilamellar vesicles were prepared from cardiolipin and produced the hexagonal II phase when dialyzed against CaCl2 or MgCl2. Upon removal of the cation by dialysis against EDTA large unilamellar vesicles were formed. The events of the transition from the lamellar to hexagonal phase and back to the lamellar phase are described.     

Selective binding of divalent cations toward heme proteins

Potential toxicity of transition metals like Hg, Cu and Cd are well known and their affinity toward proteins is of great concern. This work explores the selective nature of interactions of Cu²⁺, Hg²⁺ and Cd²⁺ with the heme proteins leghemoglobin, myoglobin and cytochrome C. The binding profiles were analyzed using absorbance spectrum and steady-state fluorescence spectroscopy. Thermodynamic parameters like enthalpy, entropy and free energy changes were derived by isothermal calorimetry and consequent binding parameters were compared for these heme proteins. Free energy (DG) values revealed Cu²⁺ binding toward myoglobin and leghemoglobin to be specific and facile in contrast to weak binding for Hg²⁺ or Cd²⁺. Time correlated single photon counting indicated significant alteration in excited state lifetimes for metal complexed myoglobin and leghemoglobin suggesting bimolecular collisions to be involved. Interestingly, none of these cations showed significant affinity for cytochrome c pointing that, presence of conserved sequences or heme group is not the only criteria for cation binding toward heme proteins, but the microenvironment of the residues or a specific folding pattern may be responsible for these differential conjugation profile. Binding of these cations may modulate the conformation and functions of these biologically important proteins.     

Interaction of complement proteins C5b-6 and C5b-7 with phospholipid vesicles: effects of phospholipid structural features

Complement components C5b-6 and C7 assemble to form C5b-7, which then interacts with membranes and commits the membrane attack complex to a target site. This protein-membrane association event was investigated to determine possible structural features that could contribute to a selective membrane attack. This system may also suggest general properties of protein-membrane insertion events. Initial binding of C5b-6 to membranes could potentially determine the site of assembly. However, binding of C5b-6 to membranes required phosphatidylglycerol or phosphatidic acid produced from egg phosphatidylcholine while binding of C5b-6 to phosphatidylcholine, phosphatidylserine, or phosphatidylinositol was undetectable. Binding to phosphatidic acid was irreversible, and the bound C5b-6 could no longer interact with C7. In contrast, C5b-7 interacted with all phospholipids tested. The rate-limiting process was the interaction of C5b-6 and C7, which displayed bimolecular properties and an activation energy of 37 kcal/mol. The C5b-7 complex showed 20-fold selectivity for small unilamellar phospholipid vesicles over large unilamellar vesicles. Vesicles carrying high negative charge densities were selected over neutral vesicles by a factor of about 5. Vesicles formed from phospholipids with short, saturated hydrocarbon side chains (dimyristoylphosphatidylcholine and dipalmitoylphosphatidylcholine) were about 5-fold less effective than those formed from phospholipids with natural fatty acid distributions. The gel vs. fluid state had little influence on C5b-7 insertion.(ABSTRACT TRUNCATED AT 250 WORDS)     

Interaction of mitochondrial malate dehydrogenase monomer with phospholipid vesicles.

The association between bovine and porcine mitochondrial malate dehydrogenase (EC 1.1.1.37) and phospholipid vesicles was investigated. At concentrations at which malate dehydrogenase exists as a dimer, entrapment within the aqueous compartment but not binding of the 14C-labelled enzyme was observed. The dissociated enzyme was labile to moderate heat and to p-chloromercuribenzoate, but in both cases inactivation was decreased by incubation with suspensions of charged phospholipid vesicles. This suggested an interaction between enzyme subunits and phospholipid, and this was confirmed by direct binding measurements and by studies that followed changes in the fluorescein-labelled enzyme. The circular-dichroism spectra of the enzyme indicated a high alpha-helix content, and suggested that a small conformational change occurred when the enzyme dissociated. Fluorescence data also suggested less-rigid molecules after dissociation. A possible mechanism, based on the flexibility of enzyme monomer and its interaction with phospholipids, by which mitochondrial matrix enzymes are specifically localized in cells, is discussed.     

Retention of aqueous contents during divalent cation-induced fusion of phospholipid vesicles

The relative kinetics of intermixing and release of liposome aqueous contents during Ca²⁺-induced membrane fusion has been investigated. Fusion was monitored by the Tb-dipicolinic acid (DPA) fluorescence assay. Release was followed by the relief of self-quenching of carboxyfluorescein or by Tb fluorescence, with essentially identical results. Fusion of large unilamellar vesicles (LUV) made of phosphatidylserine (PS) in 100 mM NaCl (pH 7.4) at 25°C was initially non-leaky, whereas the fusion of small unilamellar vesicles (SUV) was accompanied by partial release of contents. After several rounds of fusion, the internal aqueous space of the vesicles collapsed. The rate of intermixing of lipids, measured by a resonance energy transfer assay, and the rate of coalescence of aqueous contents during fusion were similar over a range of Ca²⁺ concentrations. Most of the aqueous contents were retained after the fusion of SUV (PS) in 5 mM NaCl and 1 mM Ca²⁺. LUV made of a 1:1 mixture of Bacillus subtilis cardiolipin and dioleoylphosphatidylcholine went through about two rounds of fusion in the presence of Ca²⁺ at 10°C, with complete retention of contents. Similar results were obtained with vesicles composed of phosphatidate/PS/phosphatidylethanolamine/cholesterol (1:2:3:2) in the presence of Ca²⁺ and synexin at 25°C. These results emphasize the diversity of the relative kinetics of fusion and release in different phospholipid vesicle systems under various ionic conditions, and indicate that the initial events in the fusion of LUV are in general, non-leaky.     

Interaction of factor VIII-von Willebrand Factor with phospholipid vesicles.

The interaction of Factor VIII-von Willebrand Factor with phospholipid vesicles has been studied by using sucrose-density-gradient ultracentrifugation. When purified Factor VIII-von Willebrand Factor was run alone. Factor VIII activity and Factor VIIIR-Ag sedimented together to the lower half of the tube. Addition of phosphatidylserine/phosphatidylethanolamine vesicles at concentrations above 250 microgram/ml resulted in complete separation of Factor VIII activity and Factor VIIIR-Ag, the former appearing with the phospholipid on the top of the tube and the latter sedimenting as before. This separation was obtained even in the presence of proteinase inhibitors. Activation of Factor VIII-von Willebrand Factor by thrombin resulted in formation of a slow sedimenting component containing essentially all the Factor VIII activity, whereas the Factor VIIIR-Ag sedimented towards the bottom of the tube as before. The thrombin-induced Factor VIII activity was strongly bound to phospholipid vesicles as determined by density-gradient centrifugations at various Factor VIII concentrations and low concentrations of phospholipid. Based on certain assumptions a dissociation constant of 2.5 nM was calculated, a mechanism for the formation in vivo of the Factor X-activator complex is suggested.     

Interaction of preS domains of hepatitis B virus with phospholipid vesicles

The role of preS domains of the hepatitis B virus (HBV) envelope proteins in the first steps of viral infection has been restricted to their implication in virus attachment to a putative hepatocyte receptor. In order to explore a fusion activity in these regions, we used recombinant preS domains to characterize their interaction with liposomes. Binding experiments carried out with NBD-labeled proteins indicated that preS were able to interact in a monomeric way with acidic phospholipid vesicles, being the partition coefficient similar to that described for peptides which can insert deeply into bilayers. Fluorescence depolarization of DPH-labeled vesicles confirmed the specificity for negative charged phospholipids. Upon interaction the proteins induced aggregation, lipid mixing and release of internal contents of acidic vesicles at both acid and neutral pH in a concentration-dependent manner. Taken together, all these data indicate that preS domains are able to insert into the hydrophobic core of the bilayer. Moreover, the insertion resulted in a protein conformational change which increased the helical content. Therefore all these results suggest that, besides their participation in the recognition of a cellular receptor, the preS domains could be involved in the fusion mechanism of HBV with the plasma membrane of target cells.     

Interaction of the pore-forming domain of colicin A with phospholipid vesicles

The interaction of the 20-kDa pore-forming domain of colicin A with phospholipid vesicles was investigated by gel permeation chromatography, analytical centrifugation, and electron microscopy. Under the experimental conditions of this study, this peptide was found to interact only with vesicles containing negatively charged phospholipids. It forms a well-defined disklike complex with phosphatidylglycerols with a preference for those containing 12-14 atoms of carbon in their fatty acid chain. This complex has a diameter of 120 A and is about one bilayer thick. It contains nine molecules of peptide and is formed both at acidic pH (pH 5.0) and at neutral pH (pH 7.2).     

Interaction of phospholipid vesicles with rat hepatocytes in primary monolayer culture

We studied the interaction of positively and negatively charged unilamellar and multilamellar phospholipid vesicles (liposomes) with rat-liver parenchymal cells in primary monolayer culture. Radioactive liposomal phosphatidylcholine was taken up more rapidly and to a larger extent from unilamellar than from multilamellar vesicles. No significant difference in uptake characteristics was observed between vesicles of different charge. The presence of serum greatly reduced uptake of liposomal phosphatidylcholine of both unilamellar and multilamellar vesicles. This serum effect was independent of surface charge of the vesicles. When cells were allowed to take up radioactive liposomal phospholipid and then incubated further in absence of vesicles, part of the radioactivity associated with the cells was released into the medium, most of it as water soluble degradation products. When cells were preincubated with vesicles containing horseradish peroxidase and then, after removal of the vesicles, further incubated, peroxidase activity could be demonstrated in the culture medium, part of it only after addition of Triton X-100. These observations were taken to indicate that part of the phospholipid taken up the cells represented vesicles binding to the cell surface rather than having been internalized. Vesicle-entrapped [125I]albumin was taken up by the cells and rapidly hydrolyzed as indicated by the appearance of radioactivity soluble in trichloroacetic acid within minutes after starting the incubation. No uptake of free albumin could be demonstrated. The kinetics of albumin uptake and release of trichloroacetic acid-soluble radioactivity from the cells suggest that, initially, liposomes are internalized predominantly by endocytosis, while during prolonged incubation fusion of the liposomal membrane with the plasma membrane gradually contributes more substantially to the overall uptake process. The significance of these findings is emphasized with special reference to the use of liposomes as intravenous carriers of enzymes or drugs.     

Interaction of natural and synthetic polynucleotides with liposomes in the presence of divalent cations

The formation of complexes of polynucleotides (DNA, poly A.poly U) with liposomes from egg lecithins, L-alpha-phosphatidylcholine, dimirystoyl and other lipids in the presence of divalent cations was studied by differential scanning microcalorimetry circular dichroism and turbidimetry. It was shown that the secondary structure of polynucleotides (double or triple helix) was necessary for the formation of these complexes. This structure was partially destroyed during formation of complexes. It was shown, that three main types of lipids, i.e. phosphatidylcholine, phosphatidylethanolamine and sphingomyelin participate in interactions between liposomes, polynucleotides and Mg2+.     

Insulin-dependent phosphorylation of GTP-binding proteins in phospholipid vesicles

The involvement of GTP-binding proteins (G proteins) in insulin action has been investigated in an in vitro system. Insulin receptors that have been purified by wheat germ lectin chromatography and either tyrosine-agarose chromatography, sucrose density centrifugation, or insulin-Sepharose chromatography have been co-inserted into phospholipid vesicles with different purified G proteins. The results of these studies indicate that a specific insulin-promoted phosphorylation of two G proteins, Go and Gi, can occur in these phospholipid vesicles. Bovine retinal transducin is a poor substitute for Go and Gi, being only weakly phosphorylated by the insulin receptor, and bovine brain Gs is not a substrate. The phosphorylation of Gi and Go occurs primarily on the alpha-subunits. Under optimal conditions, about one alpha o- or alpha i-subunit is phosphorylated on a tyrosine residue for every two beta-subunits of the insulin receptor, suggesting a 1:1 interaction between these G proteins and the heterotetrameric (alpha 2 beta 2) insulin receptor molecular. The inactive (GDP-bound) form of the alpha-subunits appears to be the preferred substrate, with the phosphorylation being significantly reduced in alpha o and alpha i upon the binding of guanosine 5'-O-thiotriphosphate (GTP gamma S) and completely eliminated in the pure alpha-GTP gamma S complex of transducin. The Gi and Go proteins also cause an enhancement of the insulin-stimulated receptor autophosphorylation. This enhancement is a reflection of an increased incorporation of the insulin receptor into lipid vesicles which is induced by these G proteins. Taken together these results provide evidence for the interactions of G proteins with the insulin receptor in a lipid milieu.