Polymerization of actin by positively charged liposomes.

By cosedimentation, spectrofluorimetry, and electron microscopy, we have established that actin is induced to polymerize at low salt concentrations by positively charged liposomes. This polymerization occurs only at the surface of the liposomes, and thus monomers not in direct contact with the liposome remain monomeric. The integrity of the liposome membrane is necessary to maintain actin in its polymerized state since disruption of the liposome depolymerizes actin. Actin polymerized at the surface of the liposome is organized into two filamentous structures: sheets of parallel filaments in register and a netlike organization. Spectrofluorimetric analysis with the probe N-pyrenyl-iodoacetamide shows that actin is in the F conformation, at least in the environment of the probe. However, actin assembly induced by the liposome is not accompanied by full ATP hydrolysis as observed in vitro upon addition of salts.

[Endocytosis of liposomes by Entamoeba histolytica and transport of encapsulated molecules toward the cytoplasm]. / Endocytose des liposomes par Entamoeba histolytica et transfert des molécules encapsulées vers le cytoplasme.

We have examined the interaction of a highly phagocytosing cell: Entamoeba histolytica with liposomes of different lipid compositions, and followed, by a semi-quantitative method, the intracellular fate of the entrapped molecules. Liposomes containing a small molecule, 6-carboxyfluorescein, are first phagocytosed. Then the encapsulated compound migrates from the vacuoles to the cytoplasm. Liposomes containing macromolecular substances, such as fluorescent albumin or ferritin, are also phagocytosed, but the encapsulated molecules remain within the vacuoles. We conclude that the transfer of carboxyfluorescein does not involve a fusion between liposomes and vacuoles, but more likely occurs via diffusion through membranes. The lipid composition of the liposomes does not affect phagocytosis of liposomes. In contrast, oleic acid greatly increases the transfer of carboxyfluorescein from vacuole to cytoplasm.

Modification of the ultrastructure of Entamoeba histolytica after phagocytosis of liposomes loaded with phalloidin.

The specific actin-interacting drug phalloidin has been introduced into the cytoplasm of a highly motile amoeba, Entamoeba histolytica, by a new technique: the phagocytosis of liposomes containing phalloidin. After ingestion of these liposomes, two important modifications of the ultrastructure of the amoeba were observed. First, large nodules of densely packed fine filaments are formed, which may be due to the polymerization of actin induced by the release of phalloidin within the cell's cytoplasm. Second, phalloidin induces the proliferation of ribosome crystals known as chromatoid bodies in encysted cells. This formation could be the direct consequence of the action of phalloidin on actin, where filaments form and ribosomes detach from the original oligo or polymers. However, it could also result from an unspecific toxic effect on the amoeba which, under physiological stress, starts to encyst and show multiplication of these chromatoid bodies upon encystment.

Effects of bile acids on actin polymerization in vitro.

Bile acids are major determinants of canalicular bile secretion, and there are indications that choleretic bile acids increase bile canalicular contractions, in isolated rat hepatocytes. Therefore, we examined the influence of various bile acids on the rate of actin polymerization in vitro. The free forms of cholic acid, ursodeoxycholic acid, and chenodeoxycholic acid, as well as their taurine and glycine conjugates, were incubated with purified muscle actin, at a concentration of 100-300 nmoles/mg actin. The rate of actin polymerization was measured by viscometry and the fluorescence of the pyrene probe, linked to actin. Results showed that all bile acids slow the rate of polymerization, and that the effect was dose-dependent. However, the reduction by chenodeoxycholic acid was greater than that caused by the other bile acids. The results indicate that bile acids, particularly in high concentrations interact with actin, a finding that may be related to the increased bile canalicular contractility, and altered canalicular membrane morphology, induced by choleretic bile acids.

Actin conformation is drastically altered by direct interaction with membrane lipids: a differential scanning calorimetry study.

One of the current dogmas in cytoskeleton research holds that actin filaments are attached to the cell membrane through integral membrane actin-binding proteins. We have challenged this concept, using an in vitro system composed of pure actin and liposomes, and have found that actin may also interact with membrane lipids. Differential scanning calorimetry (DSC) shows that when the actin molecule is in contact with such lipids, it undergoes a major conformational change which results in the complete disappearance of its phase transition. Conversely, DSC scans reveal that the phase transition of the membrane lipids is only weakly affected by the presence of actin. Indeed, the lipids' main transition shows only slight shifts in Tm, from 56.6 to 57 degrees C, and delta Hcal, from 10.1 to 8.8 kcal/mol. In the lipids' pretransition, Tp is shifted from 52.7 to 53.7 degrees C, and delta Hcal is shifted from 0.75 to 0.33 kcal/mol. This interaction between purified actin and membrane lipids is inhibited by high concentrations of KCl, thus indicating that the phenomenon is primarily electrostatic in nature. The ultrastructural consequences of this change in actin conformation were investigated by electron microscopy, which revealed the formation of paracrystalline arrays of actin filaments at the surface of the liposomes. We therefore propose a model in which a limited number of lipid molecules may interact with specific sites on the actin molecule, resulting in the protein's observed conformational change.

Does actin bind to membrane lipids under conditions compatible with those existing in vivo?

Using an in vitro system composed of liposomes and pure actin, we previously established that actin can interact directly with membrane lipids and suggested that this interaction may also exist in vivo. However, an important potential caveat has been brought to our attention concerning the high concentrations of lipids used in our assays. Indeed, it has been hypothesized that under our experimental conditions, divalent cations may become tightly bound to the lipid bilayers, reducing the free divalent cation concentration to non-physiological levels. The observed actin-lipid interaction has therefore been suggested to be an in vitro artifact. In order to test this hypothesis, we have measured the capture of Mg++ by DSPC liposomes under our experimental conditions. Our results show that only one Mg++ ion is captured for every 40 DSPC molecules. The resulting reduction in the free Mg++ concentration is therefore negligible in our assays. It is concluded that the actin-lipid interaction which we have previously documented indeed occurs under ionic conditions compatible with those found in vivo.

Mechanism of interaction between actin and membrane lipids: a pressure-tuning infrared spectroscopy study.

Using pressure-tuning Fourier transform infrared spectroscopy to study an in vitro system consisting of actin and distearoyl-phosphatidylcholine (DSPC) liposomes, we have determined the mechanism of interaction between actin and membrane lipids. This interaction results in a significant conformational change in actin molecules. Analysis of the amide I band of actin shows an increase in the beta-sheets to alpha-helix ratio, in random turns, and in interactions between actin monomers. In the absence of lipids, the actin molecules are denatured by pressures of 8 x 10(8) Pa and more, which give rise to a random organization of the peptide chain. However, in the presence of DSPC liposomes, pressure greater than 2 x 10(8) Pa induces a change in actin conformation, which is dominated by strongly interacting beta-sheets. As the spectra of the lipid molecules are not changed by the presence of actin, the organization of the lipid molecules in the bilayer is not affected by the protein. It is concluded from these results that this interaction of actin with membrane lipids involves very few lipid molecules. These lipid molecules may interact with actin at a few specific sites on the protein.

Virotoxins polymerize actin and induce membrane fragmentation in cytoplasmic preparations of Amoeba proteus.

Virotoxins and phalloidin are peptides that induce actin polymerization in vitro. We have compared the effect of five virotoxins and phalloidin on the ultrastructure of spread preparations of Amoeba proteus cytoplasm. Like phalloidin, the five virotoxins induce polymerization of cytoplasmic actin. Moreover, the virotoxins, but not phalloidin, induce membrane fragmentation in small spherical vesicles. We, therefore, conclude that these virotoxins may have another membrane-bound target besides actin.

Changes in molar volume and heat capacity of actin upon polymerization.

We have used densimetry and microcalorimetry to measure the changes in molar volume and heat capacity of the actin molecule during Mg(2+)-induced polymerization. Molar volume is decreased by 720 ml/mol. This result is in contradiction with previous measurements by Ikkai and Ooi [(1966) Science 152, 1756-1757], and by Swezey and Somero [(1985) Biochemistry 24, 852-860]: both of these groups reported increases in actin volume during polymerization, of 391 ml/mol and 63 ml/mol respectively. We also observed a decrease in heat capacity of about 69.5 kJ.K-1.mol-1 during polymerization. This is in agreement with the concept of conformational fluctuation of proteins proposed by Lumry and Gregory [(1989) J.Mol. Liq. 42, 113-144]whereby either ligand binding by a protein or monomer-monomer interaction decreases the protein's conformational flexibility.

Actin paracrystalline sheets formed at the surface of positively charged liposomes.

Paracrystalline aggregates of F-actin spontaneously assemble at the surface of positively charged liposomes. This single-layered paracrystalline array is made up of parallel and juxtaposed actin filaments aligned in register and showing the typical 36-nm periodicity which corresponds to the half-pitch of the double helix strand. This crystallization of pure actin results from a direct interaction between actin and positively charged lipids and does not occur with negative or neutral lipids.

Interaction of actin with positively charged phospholipids: a monolayer study.

The interactions of actin, a negatively charged cytoskeletal protein, with lipids have been studied by the monolayer technique. The lipids (either pure phosphatidylcholine or incorporating 10%, 30% or 50% of a positively charged surfactant, stearylamine) were spread at the air/water interface and actin was allowed to interact with the monolayer after injection in the subphase. The results show that at a given surface pressure, increasing the density of positive charges in the lipid monolayer causes a significant increase in the intercalation of the actin within the lipid molecules, thereby indicating that the adsorption of actin is facilitated by electrostatic interactions. However, this intercalation is only possible up to a critical surface pressure above which actin does not penetrate the lipid surface.

Effects of new triphenylethylene platinum(II) complexes on the interaction with phosphatidylcholine liposomes.

In a previous work we synthesized a class of new antineoplastic drugs by coupling a cisplatin derivative to a triphenylethylene moiety similar to the antiestrogen, tamoxifen. These drugs differ in the number of hydroxy functions on the triphenylethylene rings and in the length of the linking arm. To gain more insight into the cellular mechanism by which these new drugs act on cells, we studied, using differential scanning calorimetry, the effects of these compounds on the phase transition of membrane phospholipid (distearoyl phosphatidyl choline (DSPC)), and correlated these effects to drug cytotoxicity. The drugs without hydroxy function showed the highest cytotoxicity and induced little change on the thermogram of DSPC. Contrarily, the drugs bearing two or three hydroxy groups were less toxic, but induced important modifications of the thermogram. We suggest that the drugs with no hydroxy group enter the membrane, with the triphenylethylene moiety localized deep within the hydrophobic core of the bilayer and do not affect the cooperativity region (C2-C8). In contrast, drugs which bear hydroxy groups on the triphenylethylene rings system perturb the phospholipid molecular arrangement; this may be due either to the additional steric hindrance of the hydroxy functions in the core of the bilayer, or to their hydrophilic effect on the polar head of the lipid. In vitro, the cytotoxic effect of these drugs seems not to be related to their affinity for the estrogen receptor. We suggest that the addition of a triphenylethylene moiety to the platinum(II) complexes increases the hydrophobicity, and consequently the resulting drugs become more permeable to the membrane, particularly the non-hydroxylated triphenylethylene derivatives.

Research on the mechanism of interaction between actin and membrane lipids.

Using an in vitro system involving pure actin and liposomes, we have established that actin may interact with membrane lipids without any intermediate proteins, and that the mechanism of interaction depends upon the concentration of divalent cation. In the absence of divalent cation, actin increases membrane permeability. Low concentrations (1 mM) of divalent cation potentialize this interaction. In the presence of high divalent cation concentration, actin deposits on the surface of liposomes in a crystalline organization and reduces the membrane microviscosity as shown by the polarization of fluorescence of the DPH probe. We propose that actin interacts with lipids by hydrophobic association which is facilitated by initial electrostatic binding.

Evidence of direct interaction between actin and membrane lipids.

Actin is a protein component of the cystoskeleton and is involved in cell motility. It is believed generally that actin filaments are attached to the cell membrane through an interaction with membranous actin-binding proteins. By using an in vitro system composed of liposomes and actin, we have shown that actin may also interact directly with the phospholipids of the membrane. Actin deposited at the surface of the liposome is organized in two regular patterns: a paracrystalline sheet of parallel filaments in register, or a netlike organization. These interactions of actin with membrane lipids occur only in the presence of millimolar concentrations of Mg2+. These results suggest that the interaction of the cytoskeleton with the membrane involves, at least in part, a direct association of actin with phospholipids.

Interaction between G-actin and various types of liposomes: A 19F, 31P, and 2H nuclear magnetic resonance study.

We have investigated in the present study the interaction between G-actin and various types of liposomes, zwitterionic, positively charged, and negatively charged. To investigate at the molecular level the conformation of actin in the presence of lipids, we have selectively attached a fluorinated probe, 3-bromo-1,1,1-trifluoropropanone, to the actin cysteine residues 10, 285, and 374 and used high-resolution 19F nuclear magnetic resonance spectroscopy to investigate the probe resonances. The results indicate a change in the mobility of the 19F labels when G-actin is in the presence of positively charged liposomes made of DMPC and stearylamine and in the presence of DMPG, a negatively charged lipid. No conformational change was observed in the actin molecule in the presence of neutral liposomes. Electron micrographs of these systems reveal the formation of paracrystalline arrays of actin filaments at the surface of the positively charged liposomes, while no evidence of actin polymerization or paracrystallization was observed in the presence of DMPG. The interaction between actin and the lipid polar headgroup has also been investigated using solid-state phosphorus and deuterium NMR. The results indicate no evidence of interaction between actin and zwitterionic liposomes but show an interaction between the positively charged liposomes and a negative charge on the actin molecules. Interestingly, the negatively charged liposomes interact with a positive charge, which is most likely associated with the three residues (His-Arg-Lys) preceding the cysteine 374 residue in the protein.