Clathrin-dependent and clathrin-independent endocytosis are differentially sensitive to insertion of poly (ethylene glycol)-derivatized cholesterol in the plasma membrane.

We examined the effect of a cholesterol derivative, poly (ethylene glycol) cholesteryl ether on the structure/function of clathrin-coated pits and caveolae. Addition of the compound to cultured cells induced progressive smoothening of the surface. Markedly, when the incorporated amount exceeded 10% equivalent of the surface area, fluid pinocytosis, but not endocytosis of transferrin, became inhibited in K562 cells. In A431 cells, both clathrin-independent fluid phase uptake and the internalization of fluorescent cholera-toxin B through caveolae were inhibited with concomitant flattening of caveolae. In contrast, clathrin-mediated internalization of transferrin was not affected until the incorporated poly (ethylene glycol) cholesteryl ether exceeded 20% equivalent of the plasma membrane surface area, at which point opened clathrin-coated pits accumulated. The cells were ruptured upon further addition of poly (ethylene glycol) cholesteryl ether. We propose that the primary reason for the differential effect of poly (ethylene glycol) cholesteryl ether is that the bulk membrane phase and caveolae are both more elastic than the rigid clathrin-coated pits. We analyzed the results with the current mechanical model (Rauch and Farge, Biophys J 2000;78:3036-3047) and suggest here that the functional clathrin-lattice is much stiffer than typical phospholipid bilayers.

Poly(ethylene glycol) derivative of cholesterol reduces binding step of liposome uptake by murine macrophage-like cell line J774 and human hepatoma cell line HepG2.

Liposome uptake by HepG2 human hepatoma cells was investigated in comparison with the uptake by J774 murine macrophage-like cells. HepG2 cells accumulated liposomes (egg yolk phosphatidylcholine (EPC)/Chol; 75/25, diameter 0.2 micron) at 37 degrees C comparably to J774 macrophage-like cells. Confocal microscopic observations revealed that J774 cells internalized EPC/Chol liposomes efficiently but HepG2 cells kept most of the liposomes bound on their plasma membrane surfaces. Poly(ethylene glycol) (PEG)-coated liposomes (0.2 micron) containing poly(ethylene glycol) cholesteryl ether (PEG-Chol) avoided cellular uptake at 37 degrees C by either cell line. In both cell lines, binding of PEG-coated liposomes was lower than that of EPC/Chol liposomes when incubation was carried out at 4 degrees C. To analyze the binding process at 37 degrees C, surface-bound liposomes were removed from the cells by pronase treatment. A reduction of the amount of bound-liposomes on cell surfaces was observed in the case of PEG-coated liposomes. Therefore, PEG-coating reduces direct binding of liposomes to the cell surfaces. The presence of apolipoprotein E (apoE) increased the uptake to EPC/Chol liposomes via its receptor in both cell lines. In contrast, cellular uptake of PEG-coated liposomes was not enhanced by treatment with apoE. Therefore, while apoE-mediated liposome uptake occurs in the case of EPC/Chol liposomes, it does not occur for PEG-coated liposomes; PEG-coating also inhibits protein-mediated binding to the cells. These results further imply that elusion from liver clearance of PEG-coated liposomes is not only due to the reduction of uptake by Kupffer cells but also by hepatocytes when liposomes are small enough to go through the fenestrates of the endothelial lining.

Cholesterol derivative of poly(ethylene glycol) inhibits clathrin-independent, but not clathrin-dependent endocytosis.

The effect of poly(ethylene glycol) cholesteryl ethers (PEG(n)-Chols) with two different numbers of units (n = 50 and 200) in the hydrophilic PEG moiety on cellular endocytic activity was studied on HT-1080 cells. The amphipathic molecules were soluble in aqueous solution. When fluorescein derivatives of PEG-Chols (one fluorescein at the distal end of PEG) were incubated with the cells in culture, the cellular fluorescence was localized at the plasma membrane level and in intracellular vesicles. Fluorescence quantification indicated that for the same external concentration, twice more FPEG(50)-Chol than FPEG(200)-Chol was associated with the cells under the same conditions. Regardless of the length of PEG moiety, PEG-Chols' interaction with cells reduced the endocytic internalization of a fluid phase marker, horseradish peroxidase (HRP) depending on the cell-associated amount. In contrast, internalization of 125I-labeled epidermal growth factor (EGF) through receptor-mediated endocytosis did not change upon incubation with PEG(50)-Chol. The effect of PEG(200)-Chol was also small, since EGF internalization showed a reduction of 10-20%, while at the same concentration as much as 80% of HRP uptake was inhibited. PEG(50)-Chol did not influence the internalization of a larger ligand, 125I-transferrin (Tfn). However, in the presence of PEG(200)-Chol, the uptake of 125I-Tfn decreased remarkably, and yet, PEG(200)-Chol has no influence on the binding and internalization of a monoclonal antibody directed toward the ectodomain of the Tfn-receptor. These results suggested that incorporation of PEG-Chols in the outer monolayer of the plasma membrane specifically inhibited clathrin-independent, but not clathrin-dependent endocytosis.

The vacuolar ATPase proton pump is present on intracellular vacuoles induced by Helicobacter pylori.

Cytotoxic strains of Helicobacter pylori cause an intense vacuolar degeneration of cells, due to the enlargement of late endosomes in the presence of membrane permeant weak bases. Bafilomycins, specific inhibitors of the vacuolar-type (V-) ATPase proton pump, prevent vacuole formation. The presence of the V-ATPase on vacuolar membranes was demonstrated by immunofluorescence with a monoclonal antibody (MAb) specific for the human 116-kDa regulatory subunit. The V-ATPase co-localised with the late endosomal marker rab7 on vacuolar membranes. In contrast, the early recycling endosomal compartment was not altered by the VacA cytotoxin, although it was endowed with the V-ATPase. Endocytosis of a MAb against the 116-kDa regulatory subunit of V-ATPase blocked endosomal acidification in HeLa cells and prevented VacA action. These results indicate that selective swelling of late endosomes, due to accumulation of osmotically active weak bases driven by the V-ATPase, is essential for vacuole formation.

Wortmannin and Li+ specifically inhibit clathrin-independent endocytic internalization of bulk fluid.

Incubation of a human fibrosarcoma cell line HT-1080 in Li(+)-containing medium inhibited internalization of a fluid marker, horseradish peroxidase (HRP), by more than 80%. The ion inhibited the activity enhanced by Ca2+ or phorbol 12-myristate 13-acetate. We also found that wortmannin (WT), a potent inhibitor of phosphoinositide (PI) 3-kinase (PI 3-k), inhibited the non-stimulated and the two stimulated types of endocytosis to the same extent as Li+. In contrast, neither WT nor Li+ influenced the early internalization of transferrin (Tfn), EGF or platelet-derived growth factor. Neither targeting to early endosomes nor recycling of the once-internalized Tfn was influenced. When the cytoplasmic pH was lowered by chasing cells that had been preincubated with 25 mM NH4Cl in an amiloride-containing Na(+)-free medium, more than 90% of internalization of Tfn in HT-1080 cells was inhibited, while that of HRP was reduced by only 35%. In contrast, WT reduced the uptake of HRP by KB cells by 34%, while 60% of the activity was inhibited by the treatment for cytoplasmic acidification. Comparison of other cells i.e., A-549 and a human diploid cell line Miyajima, indicated that cells showing higher sensitivity to WT were less sensitive to low cytoplasmic pH. These results suggest that, in all the cells studied, bulk fluid is internalized either via a clathrin-independent/PI 3-k-dependent route or via a clathrin-dependent/PI 3-k-independent one, though the ratio varied among them. We also found that internalization of a mAb directed toward the 116 (100)-kDa subunit of vacuolar ATPase [OSW2; Sato and Toyama (1994) J. Cell Biol. 127, 39-53] in the fluid phase was inhibited by WT, but the antibody was still internalized in a surface-bound form. Regardless of the treatment with WT, most of the antibody was transported to endosomes that were associated with Tfn receptor. These results suggest that both internalization routes are targeted to the same early endosomal compartments.

Okadaic acid gives concentration-dependent reciprocal effects on the fluid phase endocytosis activated by Ca2+ and phorbol 12-myristate 13-acetate.

Incubation of a human fibrosarcoma cell line HT-1080 in increasing concentration of Ca2+ was found to enhance endocytic internalization of a fluid phase marker, horseradish peroxidase. At 16.8 mM Ca2+, generation of the effect required incubation for more than 45 min. The effect was reversed by removal of the excess ion for 30 min. Monitoring the intracellular concentration showed that the incubation induced a transient large Ca2+ influx followed by a recovery to 230 +/- 50 nM instead of the normal level of 83 +/- 5 nM. The activation was not inhibited by inhibitors of protein kinases nor a cAMP antagonist. In contrast, the effect was prevented by okadaic acid (OKA) at 100 nM without detectable effect on the basal activity. Fluid phase uptake by HT-1080 cells was also enhanced by phorbol 12-myristate 13-acetate (PMA). In contrast to the case with Ca2+, OKA at 100 nM did not prevent the PMA effect but further enhanced the endocytosis. The effect of OKA was concentration-dependent, as the reagent at 1 microM inhibited not only both the activation but also the basal activity. In Ca(2+)- or PMA-stimulated cells, FITC-dextran was delivered to endosomes that had been labeled with TRITC-transferrin. In contrast, following treatment with a combination of PMA and 100 mM OKA, fluid phase was internalized in vesicular compartments devoid of transferrin labeling. These results suggest that, through differential modifications of protein phosphorylation, endocytosis can be enhanced distinctively either by employing conventional receptor-bearing compartments or generating a new endosomal population.

Interference with the endosomal acidification by a monoclonal antibody directed toward the 116 (100)-kD subunit of the vacuolar type proton pump.

A monoclonal antibody (OSW2) was prepared by using human osteosarcoma cells. OSW2 was found to be directed toward the 116 (also called 100)- kD protein that uniquely associates to the vacuolar-type proton pump. The antibody specifically localized acidic membrane compartments that could be visualized with acridine orange in many types of human cells. It also reacted with the surface and was internalized along the endosomal pathway. Monitoring the endosome pH by using FITC-dextran and acridine orange suggested that the antibody interfered with low pH. Cell-free experiments indicated that the ATP-dependent acidification was inhibited in endosomes associated with OSW2. In contrast, the antibody gave little effect on the ATPase activity of the solubilized H+ pump. The internalization of OSW2 reduced infectivity of certain enveloped viruses (influenza, SFV, VSV) by 50 to 80%. Inhibition of viral fusion was directly demonstrated by monitoring the fate of octadecylrhodamine-labeled influenza virus fluorescence. These results indicate that the 116 (100)-kD protein is necessary for the control of pH. The antibody represents a novel probe for understanding the role of the endosomal compartments in cellular physiology.

Fine structure of influenza A virus observed by electron cryo-microscopy.

A rapidly frozen vitrified aqueous suspension of influenza A virus was observed by high resolution electron cryomicroscopy. The influenza particles were grouped into small (diameter < 150 nm) spherical particles with well organized interiors, large spherical ones with less internal organization, and filamentous ones. Envelopes of most of the large virus particles were phospholipid bilayers, and the chromatography fraction containing these large particles was largely devoid of viral activity. The envelopes of most of the filamentous and small spherical virus particles, on the other hand, gave a strange contrast which could be ascribed to a combination of a thin outer lipid monolayer and a 7.2 nm thick protein-containing inner layer. These latter particles represented most of the viral activity in the preparation. Densitometric traces of the near in-focus images confirmed these structural differences. Some viral envelope structures apparently intermediate between these two distinct types of membrane were also detected. A structural model of intact biologically active influenza virus particles was formulated from these results, together with computer simulations.

Subpopulations of endosomes generated at sequential stages in the endocytic pathway of asialoganglioside-containing ferrite ligands in rat liver.

Subpopulations of endosomes generated at different stages of the endocytic pathway were isolated by a high-gradient magnetic separation followed by a Percoll density gradient centrifugation. Rat livers were perfused for 5 min with asialoganglioside (ASG)-containing ferrite particles and chased at 37 degrees C. At various times after the internalization, the endocytic vesicles containing ferrite particles were isolated by the magnetic separation. Isolated fractions contained endosomes until 15-min perfusion, after which most of the particles were transported to lysosomes. The endosomal fractions isolated after the 5- or 15-min perfusions were further analyzed by 30% Percoll density gradient centrifugation. The endosomes after 5-min perfusion showed peaks around the density of 1.05 g/ml (peak I) and 1.07 g/ml (peak Is), both of which contained asialoglycoprotein receptors. In the 15-min perfusion, another peak of endosomes (peak II) was observed at the higher density of 1.09 g/ml without the receptors, in addition to peak I. These endosomes had their own characteristic proteins. Some proteins were common in the subgroups of endosomes. These results suggest that the endosome I containing the ligands and the receptors was first produced after endocytosis and, through the endosome is, was scissioned into the endosome II containing the ligands. The endosome II was then fused with primary lysosomes for proteolytic cleavage of ligands.

A novel method for isolating specific endocytic vesicles using very fine ferrite particles coated with biological ligands and the high-gradient magnetic separation technique.

We have developed a novel method for isolating specific endocytic vesicles using magnetic ligands and high-gradient magnetic separation. Ligands were prepared by coating extremely fine ferrite particles (10-20 nm) with bovine serum albumin and then conjugating asialoglycopeptides. These ligands were introduced into rat liver by perfusion at 16 or 37 degrees C, or by injection through the tail vein. The ligand particles were observed as electron-dense small grains in membrane-bound vesicles in Kupffer as well as parenchymal cells by electron microscopy. Livers were taken out, homogenized and lightly centrifuged. The supernatant was pumped into a separator glass tube filled with very fine ferritic stainless steel fibers and placed in a magnetic field of 0.9-2 T. Vesicles containing ferrite particles were collected with a high efficiency (ca. 70% of endocytosed magnetic ligands). About 70% of uptake appeared to be mediated by the asialoglycoprotein receptors. The captured vesicles were practically free from marker enzymes for plasma membranes, endoplasmic reticulum, and Golgi apparatus. Lysosomal enzyme activity of the vesicles increased with the time of perfusion at 37 degrees C but not at 16 degrees C. Protein composition of the captured vesicles was analyzed by one- and two-dimensional gel electrophoresis. The composition changed characteristically with time on perfusion at 16 and 37 degrees C. The present method provides a powerful tool to collect prelysosomal endocytic vesicles containing specific ligands and lysosomes fused with these specific endocytic vesicles.

Membrane fusion activity of reconstituted vesicles of influenza virus hemagglutinin glycoproteins.

Reconstituted vesicles of hemagglutinin glycoproteins into egg yolk phosphatidylcholine/spin-labeled phosphatidylcholine/cholesterol (molar ratio 1.6:0.4:1) were prepared by dialysis. Preparations at appropriate protein-to-lipid ratios (1:44 and 1:105 mol/mol) contained vesicles with a diameter of 100-300 nm and a high density of spikes on the surface. These vesicles showed low pH-induced membrane fusion activity. At pH 5.2 and 37 degrees C, fusion with erythrocyte membranes took place very rapidly within 1-2 min and reached a plateau at 63-66% fusion. The fusion was negligibly small at neutral pH and was induced to occur at pH values lower than 6.0. The reconstituted vesicles caused hemolysis and fusion of human erythrocyte cells in the same pH range as that of the fusion with erythrocyte membranes. The low pH-induced fusion activity of the reconstituted vesicles is essentially the same as that of the parent virus. These vesicles can be used to deliver some reagents or drugs into target cell cytoplasm via fusion at lysosomes.

Hemolytic activity of influenza virus hemagglutinin glycoproteins activated in mildly acidic environments.

Hemagglutinin (HA) glycoproteins isolated from influenza virus caused hemolysis and liposome lysis at pH less than 6.0. The pH dependence was similar to that of the parent virus. Hemagglutination and hemolysis titers of HA were comparable with those of virus. The time course of hemolysis by HA was somewhat different from that by virus. HA did not cause fusion of erythrocytes in acidic media, in contrast to virus. Both HA and virus, previously incubated at pH less than 6.0, lost their low-pH-induced hemolytic activity. Isolated HA formed rosette-like structures at neutral pH, and these aggregated in acidic media. Virus also aggregated in acidic media and its envelope became leaky to negative stain. HA previously incubated at pH less than 6.0 became susceptible to trypsin digestion. Both reversible and irreversible structural changes of HA were observed by fluorescence spectroscopy; a reversible change at a pH between neutral and 6.4 and an irreversible one at pH less than 6.0. Bromelain-released HA did not cause hemolysis and liposome lysis in acidic media. The precursor form of HA did not have hemolytic activity in acidic media. The similarity in pH dependence indicates that the structural change in HA induced at pH less than 6.0 is the cause of activation and inactivation of hemolysis, HA and virus aggregation, and trypsin susceptibility. We propose that the hydrophobic NH2-terminal segment of HA2 is exposed during the structural change and interacts with the target membranes, causing a permeability increase and leading to hemolysis and lysis. The virus-induced hemolysis can be ascribed for the most part to envelope fusion activated in acidic media.

Interaction of a peripheral protein of the erythrocyte membrane, band 4.1, with phosphatidylserine-containing liposomes and erythrocyte inside-out vesicles.

Interactions of band 4.1 with mixed phospholipid membranes [phosphatidylserine (PtdSer), phosphatidylethanolamine, phosphatidylcholine, etc.] and erythrocyte inside-out vesicles were studied. Band 4.1 showed a higher affinity to PtdSer-containing membranes. The amount of binding to PtdSer-containing liposomes was larger than that to PtdSer-lacking liposomes. The amount of binding to inside-out vesicles did not change significantly on a protease treatment of the vesicles. The amount of band 4.1 bound on inside-out vesicles decreased on PtdSer-decarboxylase treatment of the vesicles. Ca2+ acted inhibitory to the binding of band 4.1. Band 4.1 together with PtdSer-containing vesicles but not with PtdSer-lacking vesicles induced gelation of spectrin-actin copolymer solution. Ca2+ inhibited the gelation. Fluorescence energy transfer from PtdSer-containing vesicles to band 4.1 was larger than that from PtdSer-lacking vesicles. Band 4.1 caused a marked release of tempocholine from preloaded PtdSer-containing liposomes but not from PtdSer-lacking liposomes. The release was larger from liposomes containing more PtdSer. Ca2+ was inhibitory to the tempocholine release. We suggest from these results that band 4.1 provides another anchoring site for the cytoskeletal spectrin-actin network to PtdSer domains in the inner layer of erythrocyte membrane. This anchoring may be involved in functional regulation since the interaction causes the membrane permeability change that is dependent on Ca2+.

Seeding role of spectrin in polymerization of skeletal muscle actin.

The effect of spectrin on the polymerization of muscle actin has been investigated by hydrodynamic methods and electron microscopy. Spectrin markedly accelerated polymerization of actin. The effect was more easily observed in lower concentrations of KCl (e.g. 24 mM) where spontaneous polymerization was negligibly small. Similarly large acceleration was observed for polymerization in MgCl2 or CaCl2. The rate of polymerization of actin was proportionally increased with the concentration of spectrin added to a fixed concentration of action. The stationary level of specific viscosity also increased with the spectrin concentration, but at larger concentrations it became smaller. The flow birefringence and electron microscope measurements indicated that actin polymers formed under the influence of spectrin were shorter than those of control F-actin filaments. The structural viscosity and electron microscope observations suggested that the interaction between F-actin fibers was not increased by spectrin. These data strongly suggest a seeding role of spectrin in the polymerization of actin. Spectrin accelerates formation of the nuclei for polymerization. The more the nuclei are formed, the larger the number of the grown polymers are and this leads to rapid formation of shorter polymers since the amount of actin is limited. The acceleration activity was found only in freshly prepared spectrin from fresh ghosts taken from freshly drawn blood.