"Ultramicroinjection" of macromolecules or small particles into animal cells. A new technique based on virus-induced cell fusion.

A new method is described for the introduction of macromolecules and small particles into animal cells. The first step in this procedure is the trapping of particles in ghosts of human erythrocytes. This is achieved by the gradual hemolysis of erythrocytes in the presence of the particles to be trapped. The second step is the Sendai virus-induced fusion of the ghosts containing the particles with cells. By this method, ferritin and latex spheres (diameter 0.1 mum) have been "injected" into cells.     

Virus-induced fusion of human erythrocyte ghosts I. Effects of macromolecules on the final stages of the fusion reaction

Human erythrocyte ghosts prepared by hypotonic hemolysis can be fused by Sendai virus, provided that certain macromolecules (bovine serum albumin, dextran and others) are sequestered in the ghosts. Since fusion of the ghosts is dependent on intactness of the F(fusion)-glycoprotein of the virion, and since the other requirements for this reaction are also similar to those for the Sendai virus-induced fusion of intact erythrocytes, this system can be used as a model for the Sendai virus-induced cell fusion reaction. Sequestered macromolecules seem to be required for rounding of locally fused ghosts. Under low osmotic swelling conditions, such as use of ghosts sealed without macromolecules or using bovine serum albumin-loaded ghosts sealed in the presence of external macromolecules, no apparently complete cell fusion (large spherical polyghost formation) could be observed. Even under these conditions, however, occurence of local cell fusion could be demonstrated either by transfer of fluorescent-labeled albumin from one ghost to an other, or by observation of polyghost formation after osmotic swelling in the cold. Thus, final stages of the fusion reaction can be divided into local cell-cell fusion which could not be observed by phase-contrast microscopy, and rounding (i.e. formation of spherical polyghost). For the observation of fusion of ghosts, the last step seems to be important.     

Quantitative introduction of a given macromolecule into cells by fusion with erythrocyte ghosts using a fluorescence activated cell sorter.

FITC-conjugated bovine serum albumin (FITC-BSA) molecules were quantitatively introduced into human erythrocyte ghosts by gradual hemolysis. When the ghosts and L cells were fused with UV-inactivated HVJ (Sendai virus), FITC-BSA was introduced into the cytoplasm of the L cells and fluorescence could be observed inthe cells with a fluorescence microscope. A mixture of L cells and ghosts was introduced into a fluorescence activated cell sorter (FACS), which could separate the mononuclear cells on the basis of their light-scattering profile. Four distinct populations of mononuclear cells were found by fluorescence analysis. These populations were separated from the cell mixture and found to correspond to cells fused with one, two and three ghosts and unfused cells. After separation, the cells from each population could form colonies in culture. As a given macromolecule can be quantitatively introduced into erythrocyte ghosts with the FITC-BSA, after fusion of these ghosts with cells, this sorting method is useful for separating cells containing a definite number of macromolecules.     

Chemically facilitated microinjection of proteins into intact monolayers of tissue culture cells

Microinjection of physiologic quantities of macromolecules into tissue culture cells can facilitate the study of the biological effects of such macromolecules. In this communication, we describe a chemical technique which can be used to microinject proteins into monolayers of intact cells. Protein is loaded into erthrocyte ghosts, and the ghosts are then fused to the monolayer with polyethylene glycol 1000. Receipient cells can be injected with an efficiency of greater than 90% and contain an average of 3.8 X 10(6) microinjected molecules per cell. This technique circumvents certain problems encountered in virus-induced microinjection.     

Neutralization of diphtheria toxin in living cells by microinjection of antifragment A contained within resealed erythrocyte ghosts

When human erythrocytes suspended in phosphate-buffered saline (PBS) containing lgG were first dialyzed against a hypotonic solution and then dialyzed against PBS, lgG molecules were entrapped within resealed erythrocyte ghosts. The concentration of lgG inside the ghosts was about 33% of its concentration in the dialysis bag. With the aid of HVJ (Sendai virus), ghosts containing rabbit lgG antibody against fragment A of diphtheria toxin were fused with toxin-sensitive FL cells. The fused FL recipients were found to be resistant to the action of diphtheria toxin. Clones derived from the resistant recipient cells, however, became sensitive to the toxin again. Antifragment A neutralized the enzymic activity of isolated fragment A in vitro, but did not protect FL cells or rabbit skin against the complete toxin.     

Transformation of L cells with virus thymidine kinase genes introduced by red cell-mediated microinjection

Fusion of red cell ghosts containing foreign materials with cells results in the introduction of the materials into the cells (red cell-mediated microinjection). Until now, 'two-step dialysis' has mainly been used for trapping proteins in the ghosts. Large-sized materials such as DNA, however, are rarely trapped in the ghosts, since the holes in the red cell membrane caused by osmotic shock are too small for such materials to pass through. In this study, we improved the trapping technique. Some of the Hind III fragments of lambda phage DNA as well as proteins could be trapped in the ghosts when the mixture of these materials and red cells were frozen at -80 degrees C for a short period followed by quick thawing. Red cell-mediated microinjection using ghosts containing plasmid pBR322 linked with a Herpes simplex viral thymidine kinase (tk) gene brought about transformation of tk-defective L cells, the efficiency of transformation was 1 out of 20 000-60 000 cells fused with the ghosts.     

Phospholipid asymmetry in human erythrocyte ghosts

Using phospholipase digestion and the fluorescent probe merocyanine 540 the maintenance of phospholipid asymmetry in the plasma membrane of human erythrocyte ghosts was investigated. Digestion with phospholipase A2 indicated that ghosts prepared in the presence of Mg++ as the only divalent cation retained the normal phospholipid asymmetry characteristic of intact erythrocytes. These ghosts, like normal erythrocytes, also failed to stain with merocyanine 540. However, the presence of as little as 5-10 microM Ca++ during ghost preparation resulted in ghosts in which lipid asymmetry had been abolished, as indicated by phospholipase digestion. Moreover, these ghosts stained with merocyanine 540. In contrast to ghosts, intact erythrocytes treated with ionophore required millimolar levels of Ca++ ions to disrupt membrane lipid asymmetry. To discover the reason for this difference in behavior between ghosts and intact cells, ghosts were prepared from preswollen cells using only small volumes of buffer for lysis. These experiments demonstrated that as the cellular contents of erythrocytes are diluted, the asymmetric arrangement of phospholipids becomes more sensitive to disruption by Ca++.     

Microinjection of macromolecules into normal murine lymphocytes by means of cell fusion. II. Enhancement and suppression of mitogenic responses by microinjection of monoclonal anti-cyclic AMP into B lymphocytes

Reproducible methods are now available for introducing protein molecules such as antibodies into normal murine lymphocytes by fusion with protein molecule-containing erythrocyte ghosts. Monoclonal antibodies against cyclic AMP were raised by hybridoma technique and packed into erythrocyte ghosts. Then, monoclonal anti-cyclic AMP containing ghosts were fused with splenic B lymphocytes by polyethylene glycol-mediated fusion at various intervals after LPS stimulation. This method made it possible for us to quantitatively microinject antibodies into B lymphocytes. Microinjection of anti-cyclic AMP antibody molecules into lymphocytes at a very early stage of LPS stimulation resulted in a marked enhancement of DNA synthetic responses as well as increased numbers of plaque-forming cells. Intracellular cyclic AMP levels were found to be markedly decreased after microinjection of monoclonal anti-cyclic AMP, suggesting that lowering the intracellular cyclic-AMP level in the B lymphocytes at an early stage of stimulation might have induced the enhanced proliferative as well as differentiative responses to LPS. Similar enhancing effects on cell proliferation were obtained when antibodies were injected 18 hr after stimulation. Microinjection of anti-cyclic AMP at 12 hr after culture, however, inhibited the DNA synthetic responses, and induction of plaque-forming cells was suppressed when anti-cyclic AMP was injected 6 hr after LPS stimulation. The present data suggest the biphasic regulatory roles of cyclic AMP at the early stage of B lymphocyte activation. This approach may be useful in identifying regulatory molecules in B lymphocyte induced by mitogenic or antigenic stimulation.     

One molecule of diphtheria toxin fragment a introduced into a cell can kill the cell

Erythrocyte ghosts containing a known number of molecules of purified fragment A of diphtheria toxin with a constant amount of FITC-BSA as a fluorescence marker were prepared by dialyzing a mixture of erythrocytes and these substances against hypotonic solution. These substances were then introduced into diphtheria toxin-resistant mouse L cells by virus-mediated cell fusion of the cells with the ghosts, and mononuclear recipients that had fused with only one erythrocyte ghost were separated in a fluorescence-activated cell sorter (FACS) on the basis of their cell size and fluorescence intensity. After separation, the viability of cells containing known numbers of fragment A was examined by measuring colony-forming ability. The results demonstrated that a single molecule of fragment A was sufficient to kill a cell.This fact was confirmed by introduction into cells of fragment A from an immunologically related mutant toxin, CRM 176 (fragment A-176); this has a completely functional fragment B region, but in cell extracts, the enzymic activity of its fragment A is about 10 fold less than that of wild toxin. The cytotoxicity of CRM 176 is about two hundredths of that of the wild-type (Uchida, Pappenheimer and Greany, 1973). As expected, about 100–200 fold excess of fragment A-176 was needed to kill the cells.     

FUSION OF INTACT HUMAN ERYTHROCYTES AND ERYTHROCYTE GHOSTS

Sendai virus is able to induce the fusion of human erythrocytes. Bivalent cations or ATP are not essential for polyerythrocyte formation. High fusion indices were obtained when Sendai virus was added to cells incubated in the presence of both EDTA and iodoacetic acid. Human erythrocyte ghosts prepared by gradual hemolysis still retain the potential to undergo virus-induced fusion. Fusion of human red blood cells without the addition of viruses was obtained by incubation of erythrocytes at pH 10.5 in the presence of Ca⁺⁺ (40 mM) or by addition of phospholipase C Clostridium perfringens preparations to cells previously agglutinated or polylysine.     

An attempt to separate mononuclear cells fused with human red blood cell-ghosts from a cell mixture treated with HVJ (Sendai virus) using a fluorescence activated cell sorter (FACS II).

Nucleated cells (Ehrlich ascites tumor cells or L strain cells) and human red blood cells (RBC)-ghosts were mixed and fused by ultraviolet-inactivated HVJ (Sendai virus). The cell mixture was stained with FITC conjugated anti-RBC ghost antiserum and then applied to FACS II apparatus. The apparatus sorted mononuclear cells fused with RBC-ghosts from the cell mixture on the basis of both the light scattering and fluorescence profiles. When the same procedure was carried out on a mixture containing cells and intact human RBC, the cells sorted by this method were cells into which hemoglobin had been injected. The sorted cells were capable of forming colonies in culture. This sorting method may be useful for collecting cells in which macromolecules have been injected artificially by fusion of RBC-ghosts enclosing macromolecules.     

The influenza virus-induced fusion of erythrocyte ghosts does not depend on osmotic forces

The role of osmotic forces and cell swelling in the influenza virus-induced fusion of unsealed or resealed ghosts of human erythrocytes was investigated under isotonic and hypotonic conditions using a recently developed fluorescence assay (Hoekstra, D., De Boer, T., Klappe, K., Wilschut, J. (1984) Biochemistry 23, 5675-5681). The method is based on the relief of fluorescence selfquenching of the fluorescent amphiphile octadecyl rhodamine B chloride (R18) incorporated into the ghost membrane as occurs when labeled membranes fuse with unlabeled membranes. No effect neither of the external osmotic pressure nor of cell swelling on virally mediated ghost fusion was established. Influenza virus fused unsealed ghosts as effectively as resealed ghosts. It is concluded that neither osmotic forces nor osmotic swelling of cells is necessary for virus-induced cell fusion. This is supported by microscopic observations of virus-induced fusion of intact erythrocytes in hypotonic and hypertonic media. A disruption of the spectrin-actin network did not cause an enhanced cell fusion at acidic pH of about 5 or any fusion at pH 7.4.     

Ghost-mediated transfer of human hypoxanthine-guanine phosphoribosyl transferase into deficient Chinese hamster ovary cells by means of polyethylene glycol-induced fusion

Chinese hamster ovary (CHO) cells deficient in hypoxanthine-guanine phosphoribosyl transferase (HGPRT) have been fused by means of polyethylene glycol (PEG) with erythrocyte ghosts loaded with crude extracts of human HGPRT. When a ratio of 100 loaded ghosts per deficient CHO cell was used about 80% of the cells showed grains above the background. Mononucleated cells which comprised about 81% of the injected cells contained an average number of grains of 36 ± 1 compared with 96 ± 2 found in mononucleated wild-type cells fused and labelled under the same conditions. Cell viability was not greatly affected after injection as nearly 85% of the total cell population excluded trypan blue 22 h after fusion and at least 90% of the mononucleated cells divided within 30 h after fusion.     

Low concentrations of macromolecular solutes significantly affect electrofusion yield in erythrocyte ghosts

Electrofusion yields in rabbit erythrocyte ghosts containing various amounts of hemoglobin, bovine serum albumin, or dextran at low concentrations were measured as a function of pulse field strength and pulse decay half-time. The presence of any of the macromolecules in low concentrations caused fusion yields to be significantly higher than when the ghosts were white (i.e., containing only buffer). The fusion yield enhancement was also critically dependent on the parameters of the electric field pulse. The fusion yield was also significantly affected by small changes in the concentration of hemoglobin when it was present outside the ghost membranes in the suspension buffer.     

Degradation of microinjected proteins: Effects of lysosomotropic agents and inhibitors of autophagy

HeLa cells, injected with radioiodinated proteins by fusion with RBC ghosts, were exposed to inhibitors of lysosomal proteolysis and autophagy. The degradation of injected [¹²⁵I]bovine serum albumin (BSA) was unaffected by chloroquine, NH₄Cl, nocodazole, colcemid, puromycin, cycloheximide, or enucleation. Although degradation of [¹²⁵I]lactate dehydrogenase (LDH) and [¹²⁵I]pyruvate kinase (PK) was inhibited one-third by chloroquine or ammonia, their degradation was unaffected by the other compounds. In contrast, enhanced degradation of ¹²⁵I-PK resulting from depriving injected HeLa cells of amino acids and serum was inhibited 70% by colcemid and abolished by chloroquine or ammonia. Similarly, degradation of [¹⁴C]sucrose-labeled BSA-polylysine conjugates that entered HeLa cells by endocytosis was inhibited as much as 80% by chloroquine and ammonia. Sensitivity of both enhanced proteolysis and degradation of exogenous proteins to ammonia or chloroquine indicates they are effective inhibitors of lysosomal proteolysis in HeLa cells. Failure of ammonia or chloroquine to inhibit degradation of injected ¹²⁵I-BSA and the modest inhibition of degradation of injected ¹²⁵I-LDH or ¹²⁵I-PK indicates that virtually all BSA molecules and most PK or LDH molecules are degraded by a nonlysosomal proteolytic system. Components of this degradative system are present in vast excess or are long lived, since inhibition of protein synthesis for 20 hr had no effect on the degradation of injected proteins.     

The interaction of 1-fluoro-2,4-dinitrobenzene with amino-phospholipids in membranes of intact erythrocytes,modified erythrocytes,and erythrocytes ghosts

Summary 1-Fluoro-2,4-dinitrobenzene (FDNB) has been used to study the availability of amino-containing phospholipids in erythrocyte membranes and ghosts in an aqueous, isotonic medium. It was found that the addition of bovine serum albumin (BSA) to the medium protects the cells from cation leak and protects some of the amino-phospholipids from reacting with the probe. In isotonic medium without BSA, 46% of the phosphatidylethanolamine and 12% of the phosphatidylserine of erythrocytes and 73% and 21% of these respective lipids of ghosts react with the probe. In the presence of 70 m BSA, 31% of phosphatidylethanolamine and 6.5% of phosphatidylserine of erythrocytes and 59% and 16% of these respective lipids of ghosts react with the probe. The labeling of these lipids does not change under conditions of varying tonicity, or after treatment of erythrocytes with pronase or lysolecithin. The data suggest that 46% of phosphatidylethanolamine and 12% of phosphatidylserine of the erythrocyte membrane are free in a lipid bilayer; 27% and 9% of these respective lipids are loosely bound to proteins which are lost during the preparation of ghosts and 27% of the phosphatidylethanolamine and 79% of the phosphatidylserine are tightly bound to core proteins which remain part of the erythrocyte membrane even after hemolysis.     

Osmotic forces in artificially induced cell fusion

The importance of cell swelling in the fusion of erythrocytes by three different chemical treatments has been investigated with cells that were cytoplasmically labelled with 6-carboxyfluorescein. Hen erythrocytes, which had been pre-incubated with ionophore A23187 and 5 mM Ca2+ to cause a proteolytic breakdown of the membrane skeleton, were induced to fuse by applying an osmotic shock. Human erythrocytes that had been incubated in an isotonic salt/buffer solution, which was progressively diluted and which contained 0.5 mM La3+ to minimise cell lysis, were also fused. In addition, the fusion of human erythrocytes by 40% poly(ethylene glycol) began only when the poly(ethylene glycol) was diluted, and it mostly occurred when the diluted polymer solution was subsequently replaced by isotonic buffer. In related experiments, the effect of an osmotic gradient on electrically induced cell fusion has been studied. Human erythrocytes in 150 mM erythritol fused more readily than less swollen cells in 200-400 mM erythritol when subjected to a 20 microseconds pulse of 3.5 kV X cm-1, indicating that the extent of cell fusion induced by the breakdown pulse is governed by the combined electrical-compressive and osmotic forces. Since osmotic phenomena are already known to be important in exocytosis, we suggest that these observations on cell fusion indicate that osmotic forces may provide the driving force for many membrane fusion reactions in biological systems.     

The Effect of PS Content on the Ability of Natural Membranes to Fuse with Positively Charged Liposomes and Lipoplexes

Supramolecular aggregates containing cationic lipids have been widely used as transfection mediators due to their ability to interact with negatively charged DNA molecules and biological membranes. First steps of the process leading to transfection are partly electrostatic, partly hydrophobic interactions of liposomes/lipoplexes with cell and/or endosomal membrane. Negatively charged compounds of biological membranes, namely glycolipids, glycoproteins and phosphatidylserine (PS), are responsible for such events as adsorption, hemifusion, fusion, poration and destabilization of natural membranes upon contact with cationic liposomes/lipoplexes. The present communication describes the dependence of interaction of cationic liposomes with natural and artificial membranes on the negative charge of the target membrane, charges which in most cases were generated by charging the PS content or its exposure. The model for the target membranes were liposomes of variable content of PS or PG (phosphatidylglycerol) and erythrocyte membranes in which the PS and other anionic compound content/exposure was modified in several ways. Membranes of increased anionic phospholipid content displayed increased fusion with DOTAP (1,2-dioleoyl-3-trimethylammoniumpropane) liposomes, while erythrocyte membranes partly depleted of glycocalix, its sialic acid, in particular, showed a decreased fusion ability. The role of the anionic component is also supported by the fact that erythrocyte membrane inside-out vesicles fused easily with cationic liposomes. The data obtained on erythrocyte ghosts of normal and disrupted asymmetry, in particular, those obtained in the presence of Ca²⁺, indicate the role of lipid flip-flop movement catalyzed by scramblase. The ATP-depletion of erythrocytes also induced an increased sensitivity to hemoglobin leakage upon interactions with DOTAP liposomes. Calcein leakage from anionic liposomes incubated with DOTAP liposomes was also dependent on surface charge of the target membranes. In all experiments with the asymmetric membranes the fusion level markedly increased with an increase of temperature, which supports the role of membrane lipid mobility. The decrease in positive charge by binding of plasmid DNA and the increase in ionic strength decreased the ability of DOTAP liposomes/lipoplexes to fuse with erythrocyte ghosts. Lower pH promotes fusion between erythrocyte ghosts and DOTAP liposomes and lipoplexes. The obtained results indicate that electrostatic interactions together with increased mobility of membrane lipids and susceptibility to form structures of negative curvature play a major role in the fusion of DOTAP liposomes with natural and artificial membranes.     

Interaction of photosynthetic pigments with various organic solvents. Magnetic circular dichroism approach and application to chlorosomes

Reduction of extracellular ferricyanide by intact cells reflects the activity of an as yet unidentified trans-plasma membrane oxidoreductase. In human erythrocytes, this activity was found to be limited by the ability of the cells to recycle intracellular ascorbic acid, its primary trans-membrane electron donor. Ascorbate-dependent ferricyanide reduction by erythrocytes was partially inhibited by reaction of one or more cell-surface sulfhydryls with p-chloromercuribenzene sulfonic acid, an effect that persisted in resealed ghosts prepared from such treated cells. However, treatment of intact cells with the sulfhydryl reagent had no effect on NADH-dependent ferricyanide or ferricytochrome c reductase activities of open ghosts prepared from treated cells. When cytosol-free ghosts were resealed to contain trypsin or pronase, ascorbate-dependent reduction of extravesicular ferricyanide was doubled, whereas NADH-dependent ferricyanide and ferricytochrome c reduction were decreased by proteolytic digestion. The trans-membrane ascorbate-dependent activity was also found to be inhibited by reaction of sulfhydryls on its cytoplasmic face. These results show that the trans-membrane ferricyanide oxidoreductase is limited by the ability of erythrocytes to recycle intracellular ascorbate, that it does not involve the endofacial NADH-dependent cytochrome b(5) reductase system, and that it is a trans-membrane protein that contains sensitive sulfhydryl groups on both membrane faces.     

Fusion of rat erythrocytes by membrane-mobility agent A2C depends on membrane proteolysis by a cytoplasmic calpain

The membrane-mobility agent 2-(2-methoxyethoxy)ethyl-cis-8-(2-octylcyclopropyl)octanoate (A2C) promotes fusion of rat, but not of human, erythrocytes. The difference in fusibility was shown to be correlated with membrane proteolysis, a process induced by Ca2+ in the rat erythrocytes or hemolysate-loaded ghosts, but not in the human cell. Membrane proteolysis is necessary but not sufficient for fusion. Fusion requires both Ca2+ and A2C [Kosower, N. S., Glaser, T. and Kosower, E. M. (1983) Proc. Natl Acad. Sci USA 80, 7542-7546]. Membrane proteolysis (Ca2+-dependent) and fusion (Ca2+ and A2C-dependent) requires a Ca2+-activated cytoplasmic thiol protease, as shown by the following observations. In intact rat erythrocytes, proteolysis and fusion are prevented by thiol alkylation and by inhibitors of Ca2+-dependent thiol proteases. Inhibitors to other proteases have no effect. Erythrocyte ghosts undergo proteolysis and fusion only when loaded with non-inhibited hemolysate, irrespective of membrane status (native or alkylated membrane). A partially purified cytosolic enzyme, identified as calpain I, promotes proteolysis in rat erythrocyte ghosts. A2C induces fusion only in such calpain-treated ghosts.