Endosomes and toxin translocation

In this review we discuss data obtained by our group regarding the entry of toxins, especially ricin, diphtheria toxin (DT) and Pseudomonas exotoxin A (PE) into animal cells. We studied the translocation process of these toxins using endosomes purified from lymphocytes. This process is rate-limiting for toxicity and enables these toxins to reach the cytosol where they will inactivate the protein synthesis system and kill the cell. We could show that each of these toxins uses a different strategy to cross the endosome membrane. Whereas ricin transmembrane transport only relies on cytosolic ATP hydrolysis, PE first requires exposure to the low endosomal pH (pH-6), presumably to insert into the endosome membrane, before being translocated via a process which also requires cytosolic ATP hydrolysis. DT translocation is directly triggered and energized by the endosome-cytosol pH gradient. Using conjugates with dihydrofolate reductase we could indirectly show that ricin and PE require unfolding for translocation. A deletion approach enabled to produce a more cytotoxic PE mutant by increasing its translocation activity.     

Energy requirements for diphtheria toxin translocation are coupled to the maintenance of a plasma membrane potential and a proton gradient

Translocation of diphtheria toxin (DT) or ricin to the cytosol is the rate-limiting step responsible for (pseudo) first-order decline in protein synthesis observed in intoxicated cell populations. The requirements for energy utilization in the translocation of both toxins are examined by perturbing the intoxication during this period of protein synthesis decline. Translocation of either toxin is blocked at 4 degrees C and requires energy. Ricin translocation is tightly coupled to ATP hydrolysis with no involvement of membrane potential. Cell depolarization slows the rate of DT translocation but does not block completely. Elimination of transmembrane pH gradients alone does not affect DT translocation; however, in combination with depolarization, translocation is blocked virtually completely. Energy requirements for DT intoxication are mediated by establishing a plasma membrane potential and a pH gradient across some cellular membrane. It is proposed that a postendocytotic vesicle containing processed DT fuses with the plasma membrane. Either component of the proton motive force across the plasma membrane then drives DT translocation. Ricin apparently utilizes a different energy coupling mechanism at a different intracellular site, thus demonstrating toxin specificity in the translocation mechanism.     

Brefeldin-A enhancement of ricin A-chain immunotoxins and blockade of intact ricin, modeccin, and abrin

After binding, the protein toxins ricin, abrin, and modeccin are endocytosed and processed through the cell's vesicular system in a poorly understood fashion, prior to translocation to the cytosol. The role of the Golgi apparatus in toxin processing was studied using brefeldin-A (BFA), a fungal metabolite which blocks Golgi function. At concentrations that inhibit secretion of interleukin-2 (IL-2), BFA blocks ricin, modeccin, and abrin intoxication of a lymphocyte derived cell line (Jurkat). Paradoxically, BFA enhances the toxicity of two ricin A-chain immunotoxins targeted against distinct cell surface determinants. BFA concentrations which are optimal for immunotoxin enhancement are below those needed to affect ricin intoxication or IL-2 secretion. BFA blockade of ricin does not involve effects on ricin endocytosis, toxin translocation to the cytosol, or the enzymatic activity of toxin A-chain. In contrast, BFA has no effect on immunotoxin processing but does enhance the immunotoxin translocation step. It is concluded that: 1) intact Golgi function is required for holotoxin processing. 2) Intact Golgi function is not required for holotoxin translocation. 3) Golgi function is tightly linked to immunotoxin translocation. 4) BFA has effects on vesicular routing in addition to the block of Golgi function in secretion which has been reported.     

Disruption of the Golgi apparatus by brefeldin A inhibits the cytotoxicity of ricin, modeccin, and Pseudomonas toxin

We have studied the cytotoxicity of ricin in cells treated with brefeldin A (BFA), which dramatically disrupts the structure of the Golgi apparatus causing Golgi content and membrane to redistribute to the ER. BFA inhibits the cytotoxicity of ricin in Chinese hamster ovary, normal rat kidney, and Vero cells and abolishes the enhancement of ricin cytotoxicity by NH4Cl, nigericin, swainsonine, and tunicamycin or by a mutation in endosomal acidification. BFA protects cells from the cytotoxicities of modeccin and Pseudomonas toxin, but has no effect on the intoxication by diphtheria toxin. Pretreatment of BFA does not protect cells from ricin treatment in the absence of BFA. Our results suggest that ricin, modeccin, and Pseudomonas toxin share a common pathway of intracellular transport from endosomes to the Golgi region where they are released into the cytosol. In contrast, the lack of protection of Vero cells from diphtheria toxin by BFA indicates that diphtheria toxin is released from acidified endosomes without involving the Golgi region.     

Enhancement of cytotoxicities of ricin and Pseudomonas toxin in Chinese hamster ovary cells by nigericin.

Nigericin and monensin, ionophores for Na+ and K+, have been found to enhance the cytotoxicities of abrin, ricin, and Pseudomonas aeruginosa exotoxin A in Chinese hamster ovary (CHO) cells. They do not affect the cytotoxicity of diphtheria toxin in the same cell line. Maximal sensitization of the CHO cells toward ricin and Pseudomonas toxin requires preculture of CHO cells in the presence of nigericin. Inhibition of protein synthesis in CHO cells by ricin or Pseudomonas toxin is also enhanced by preculture of CHO cells in the presence of nigericin. These results suggest a common step in the intoxication process of ricin and Pseudomonas toxin, the rate of which is facilitated by pretreatment with nigericin. This step is, however, not shared by the intoxication of CHO cells with diphtheria toxin.     

A class of mutant CHO cells resistant to cholera toxin rapidly degrades the catalytic polypeptide of cholera toxin and exhibits increased endoplasmic reticulum-associated degradation

After binding to the eukaryotic cell surface, cholera toxin undergoes retrograde transport to the endoplasmic reticulum. The catalytic A1 polypeptide of cholera toxin (CTA1) then crosses the endoplasmic reticulum membrane and enters the cytosol in a process that may involve the quality control mechanism known as endoplasmic reticulum-associated degradation. Other toxins such as Pseudomonas exotoxin A and ricin are also thought to exploit endoplasmic reticulum-associated degradation for entry into the cytosol. To test this model, we mutagenized Chinese hamster ovary cells and selected clones that survived a prolonged coincubation with Pseudomonas exotoxin A and ricin. These lethal endoplasmic reticulum-translocating toxins bind different surface receptors and target different cytosolic substrates, so resistance to both would likely result from disruption of a shared trafficking or translocation event. Here we characterize two Pseudomonas exotoxin A/ricin-resistant clones that exhibited increased endoplasmic reticulum-associated degradation. Both clones acquired the following unselected traits: (i) resistance to cholera toxin; (ii) increased degradation of an endoplasmic reticulum-localized CTA1 construct; (iii) increased degradation of an established endoplasmic reticulum-associated degradation substrate, the Z variant of alpha1-antitrypsin (alpha1AT-Z); and (iv) reduced secretion of both alpha1AT-Z and the transport-competent protein alpha1AT-M. Proteosome inhibition partially rescued the alpha1AT-M secretion deficiencies. However, the mutant clones did not exhibit increased proteosomal activity against cytosolic proteins, including a second CTA1 construct that was expressed in the cytosol rather than in the endoplasmic reticulum. These results suggested that accelerated endoplasmic reticulum-associated degradation in the mutant clones produced a cholera toxin/Pseudomonas exotoxin A/ricin-resistant phenotype by increasing the coupling efficiency between toxin translocation and toxin degradation.     

Role of lipids in the retrograde pathway of ricin intoxication

The plant toxin ricin binds to both glycosphingolipids and glycoproteins with terminal galactose and is transported to the Golgi apparatus in a cholesterol-dependent manner. To explore the question of whether glycosphingolipid binding of ricin or glycosphingolipid synthesis is essential for transport of ricin from the plasma membrane to the Golgi apparatus, retrogradely to the endoplasmic reticulum or for translocation of the toxin to the cytosol, we have investigated the effect of ricin and the intracellular transport of this toxin in a glycosphingolipid-deficient mouse melanoma cell line (GM95), in the same cell line transfected with ceramide glucosyltransferase to restore glycosphingolipid synthesis (GM95-CGlcT-KKVK) and in the parental cell line (MEB4). Ricin transport to the Golgi apparatus was monitored by quantifying sulfation of a modified ricin molecule, and toxicity was studied by measuring protein synthesis. The data reveal that ricin is transported retrogradely to the Golgi apparatus and to the endoplasmic reticulum and translocated to the cytosol equally well and apparently at the same rate in cells with and without glycosphingolipids. Importantly cholesterol depletion reduced endosome to Golgi transport of ricin even in cells without glycosphingolipids, demonstrating that cholesterol is required for Golgi transport of ricin bound to glycoproteins. The rate of retrograde transport of ricin was increased strongly by monensin and the lag time for intoxication was reduced both in cells with and in those without glycosphingolipids. In conclusion, neither glycosphingolipid synthesis nor binding of ricin to glycosphingolipids is essential for cholesterol-dependent retrograde transport of ricin. Binding of ricin to glycoproteins is sufficient for all transport steps required for ricin intoxication.     

Entry of lethal doses of abrin, ricin and modeccin into the cytosol of HeLa cells

In attempts to assess how many molecules of the toxic lectins abrin, ricin and modeccin are needed in the cytosol to kill HeLa cells the effect of these toxins on protein synthesis and plating efficiency was studied. The incubation time of the cells after a 1 h exposure to the toxins influenced strongly the extent of inhibition of protein synthesis. The full toxic effect was expressed about 20 h of incubation after the exposure. On further incubation, protein synthesis again increased at a rate comparable to that in the control cells. After exposure to increasing concentrations of toxins the inhibition of cellular protein synthesis measured after 20 h showed excellent agreement with the inhibition of plating efficiency, indicating that the inhibition of protein synthesis can be used as a measure of cell killing. The inhibition of protein synthesis by toxins was found to follow first order kinetics, indicating that the cells are killed by an all- or none-effect. Autoradiographic studies indicated that after exposure to intermediate toxin concentrations protein synthesis was completely abolished in some cells, whereas it appeared to proceed at a normal rate in the remaining cells. The results provide evidence that penetration of one molecule of abrin, ricin or modeccin into cytosol is lethal to HeLa cells and that the efficiency of toxin entry into the cytoplasm is very low compared to the rate of bulk toxin uptake.     

Toxin resistance and reduced secretion in a mouse L-cell mutant defective in herpes virus propagation.

The mouse L-cell mutant gro29 was selected originally for its inability to propagate herpes simplex virus; it shows severe defects in virus egress and the transport and processing of viral glycoproteins after infection. In this report, we show that uninfected gro29 cells display pleiotropic changes in protein secretion, oligosaccharide processing, and sensitivity to the toxins ricin and modeccin. Specifically, the rate of secretion of a nonglycosylated protein, human growth hormone, was reduced 70% in gro29 cells compared with the parental L cells. A direct measurement of the transport capacity of Golgi membranes in a cell-free assay suggests that gro29 cells contain less functional Golgi than parental cells. Despite this deficiency, N-linked oligosaccharides were processed efficiently in mutant cells, although there were differences in the structure of the mature forms. Lectin intoxication assays revealed that gro29 cells were cross-resistant to killing by the cytotoxic lectins ricin and modeccin, but not to wheat germ agglutinin, Ricinus communis agglutinin RCA120, or leucoagglutinin. Fluorescence labeling using fluorescein-conjugated lectins showed that uninfected gro29 cells expressed relatively few ricin-binding molecules, suggesting a possible mechanism for toxin resistance. These studies provide evidence that the processes of protein secretion, lectin intoxication, and herpes virus maturation and egress may share a common cellular component.     

Enhancement of cytotoxicity of modeccin by nigericin in modeccin-resistant mutant cell lines

We have isolated a Chinese hamster ovary cell mutant (DMPR-2) simultaneously resistant to diphtheria toxin and modeccin. In addition to the increased resistance to these two toxins used in the selection, this mutant is more resistant to Pseudomonas toxin and hypersensitive to ricin than the parental cell line. In contrast to the wild-type cells in which nigericin protects cells from modeccin, the cytotoxicity of modeccin in the DMPR-2 mutant is enhanced by nigericin. We have also studied the effects of nigericin and NH4Cl on the cytotoxicity of modeccin in a modeccin-resistant mutant of HeLa cells (ModRI). The cytotoxicity of modeccin is enhanced by nigericin in ModRI mutant cells, in contrast to the protection of modeccin cytotoxicity by nigericin in the parental HeLa cells. Our results suggest that modeccin can reach the cytosol of mammalian cells by two distinct routes; the major route requires endosomal acidification and the minor route is activated by nigericin.     

Saporin and ricin A chain follow different intracellular routes to enter the cytosol of intoxicated cells

Several protein toxins, such as the potent plant toxin ricin, enter mammalian cells by endocytosis and undergo retrograde transport via the Golgi complex to reach the endoplasmic reticulum (ER). In this compartment the catalytic moieties exploit the ER-associated degradation (ERAD) pathway to reach their cytosolic targets. Bacterial toxins such as cholera toxin or Pseudomonas exotoxin A carry KDEL or KDEL-like C-terminal tetrapeptides for efficient delivery to the ER. Chimeric toxins containing monomeric plant ribosome-inactivating proteins linked to various targeting moieties are highly cytotoxic, but it remains unclear how these molecules travel within the target cell to reach cytosolic ribosomes. We investigated the intracellular pathways of saporin, a monomeric plant ribosome-inactivating protein that can enter cells by receptor-mediated endocytosis. Saporin toxicity was not affected by treatment with Brefeldin A or chloroquine, indicating that this toxin follows a Golgi-independent pathway to the cytosol and does not require a low pH for membrane translocation. In intoxicated Vero or HeLa cells, ricin but not saporin could be clearly visualized in the Golgi complex using immunofluorescence. The saporin signal was not evident in the Golgi, but was found to partially overlap with that of a late endosome/lysosome marker. Consistently, the toxicities of saporin or saporin-based targeted chimeric polypeptides were not enhanced by the addition of ER retrieval sequences. Thus, the intracellular movement of saporin differs from that followed by ricin and other protein toxins that rely on Golgi-mediated retrograde transport to reach their retrotranslocation site.     

Lipid requirements for entry of protein toxins into cells

The plant toxin ricin and the bacterial toxin Shiga toxin both belong to a group of protein toxins having one moiety that binds to the cell surface, and another, enzymatically active moiety, that enters the cytosol and inhibits protein synthesis by inactivating ribosomes. Both toxins travel all the way from the cell surface to endosomes, the Golgi apparatus and the ER before the ribosome-inactivating moiety enters the cytosol. Shiga toxin binds to the neutral glycosphingolipid Gb3 at the cell surface and is therefore dependent on this lipid for transport into the cells, whereas ricin binds both glycoproteins and glycolipids with terminal galactose. The different steps of transport used by these toxins have specific requirements for lipid species, and with the recent developments in mass spectrometry analysis of lipids and microscopical and biochemical dissection of transport in cells, we are starting to see the complexity of endocytosis and intracellular transport. In this article we describe lipid requirements and the consequences of lipid changes for the entry and intoxication with ricin and Shiga toxin. These toxins can be a threat to human health, but can also be exploited for diagnosis and therapy, and have proven valuable as tools to study intracellular transport.     

A comparative study of ricin and diphtheria toxin-antibody-conjugate kinetics on protein synthesis inactivation

Kinetic data on toxin and antibody-toxin-conjugate inactivation of protein synthesis have been used to assess the variables which affect the transport of these toxins into the cytosol compartment. First-order inactivation rate constants of protein synthesis (ki) are compared under conditions of known receptor occupancy. The effect of inclusion of toxin B chains, both homologous and heterologous, in antibody-toxin conjugates is observed, and factors which affect toxin lag periods are studied. The results show that the inclusion of B chains in conjugates increases ki values 3-10-fold, but only if the B chain is homologous with the A chain. In spite of the augmentation of antibody-toxin-conjugate ki values by homologous toxin B chain, these ki values are only 1/20 those observed with unmodified toxins on sensitive cells. A further difference noted between toxins and antibody-toxin conjugates is the presence of a dose-dependent lag when toxins, but not antibody-toxin conjugates, effect sensitive cell types. This lag period for ricin can be shortened by alkalinizing the cell medium. The kinetic data can be fit by assuming a processing step interposed between the binding of ricin to surface receptors and the interaction of the A chain with ribosomes which is first-order in toxin concentration and pH-dependent. The time constant of this event is reflected in the dose-dependent lag period. It is proposed that antibody-toxin conjugates do not participate in this processing event and therefore fail to achieve the high entry levels exhibited by unmodified toxins.     

Properties and action mechanism of the toxic lectin modeccin: Interaction with cell lines resistant to modeccin,abrin, and ricin

The toxic lectin modeccin, which inhibits protein synthesis in eukaryotic cells, is cleaved upon treatment with 2-mercaptoethanol into two peptide chains which move in polyacrylamide gels at rates corresponding to molecular weights 28,000 and 38,000. After reduction, the toxin loses its effect on cells, while its ability to inhibit cell-free protein synthesis increases. Like abrin and ricin it inhibits protein synthesis by inactivating the 60S ribosomal subunits. Modeccin binds to surface receptors containing terminal galactose residues. Competition experiments with various glycoproteins indicate that the modeccin receptors are different from the abrin receptors. In addition, they were present on HeLa cells in much smaller numbers. Moreover, mutant lines resistant to abrin and ricin were not resistant to modeccin and vice-versa. The toxin resistance of various mutant cell lines could not be accounted for by a reduced number of binding sites on cells. The data are consistent with the view that the cells possesss different populations of binding sites with differences in ability to facilitate the uptake of the toxins and that in the resistant lines the most active receptors have been reduced or eliminated.     

A toxin-resistant mouse L-cell mutant defective in protein transport along the secretory pathway.

Using methods designed for isolation of mutants defective in receptor-mediated endocytosis, a novel L-cell mutant was obtained that exhibits resistance to three different protein toxins as well as alterations in secretion. This mutant, LEFIC, is resistant to modeccin, Pseudomonas exotoxin, and ricin. These toxins, which enter the cytoplasm via receptor-mediated endocytosis, are thought to penetrate into cells at the level of late endosomes or the trans Golgi network. Early endosomal acidification appears to be normal in the mutant based on its accumulation of iron from transferrin and its sensitivity to diphtheria toxin A chain-transferrin conjugate. Within the secretory pathway two delays in transport of vesicular stomatitis virus (VSV) G protein were observed in LEFIC: a 20-30 min delay in acquisition of Endo H resistance and a 1-2 hr delay in appearance of newly synthesized G protein on the cell surface. Movement of endogenous proteins along the secretory pathway was also affected in LEFIC. Fibronectin secretion was delayed by 15 min, and membrane proteins were delayed in arrival at the cell surface. The phenotype of LEFIC is consistent with a defect in a component or compartment shared by both the late endocytic and constitutive secretory pathways.     

Response of cultured mammalian cells to the exotoxins of Pseudomonas aeruginosa and Corynebacterium diphtheriae: differential cytotoxicity.

The sensitivities of 21 mammalian cell lines to the exotoxins of Pseudomonas aeruginosa and Corynebacterium diphtheriae were measured. Each line exhibited 1-4 log differences in sensitivities to the two toxins. No species-specific sensitivities were noted for Pseudomonas exotoxin while diphtheria exotoxin was most potent in cells of monkey origin, followed by human and hamster cells. Rat- and mouse-derived cell lines were very insensitive to diphtheria exotoxin. The rates of cellular intoxication by both toxins exhibited apparent first-order kinetics and were indistinguishable from one another when equipotent doses were used. Our preparation of diphtheria exotoxin appeared to have a slightly higher ADP-ribosylating efficiency than did Pseudomonas toxin. However, neither toxin exhibited cell line-specific differences in ribosylating efficiencies which could have explained the wide range in potencies for intact cells. Our results suggest that there are significant differences in the mechanisms of cellular intoxication by Pseudomonas and diphtheria exotoxins and that these differences probably exist in the attachment or internalization stages of toxin action.     

Specific efflux of glutathione from the basolateral membrane domain in polarized MDCK cells during ricin-induced apoptosis.

Although the depletion of reduced glutathione (GSH) has been observed in a variety of apoptotic systems, little is known about the mechanism of GSH depletion. In this study we used polarized MDCK cells to study the GSH flux during ricin-induced apoptosis. Here we report that the specific accumulation of GSH occurred in the basolateral medium during ricin treatment with similar kinetics to in apoptotic changes such as an increase in caspase-3 like activity and DNA fragmentation, while there was no significant increase in the GSH level in apical medium. These results suggest that GSH efflux occurred through a GSH-specific channel or transporter located in the basolateral membrane domain of polarized MDCK cells undergoing apoptosis. Treatment with other protein toxins such as modeccin, Pseudomonas toxin, and diphtheria toxin, which can induce apoptotic cell death, also resulted in selective GSH efflux from the basolateral side. Thus, GSH efflux through a specific transporter may be a common step of apoptosis induced by these toxins, while these toxins have different intoxication mechanisms leading to protein synthesis inhibition. Pretreatment of cells with Z-Asp-CH(2)-DCB, a caspase family inhibitor, inhibited ricin-induced basolateral GSH efflux as well as DNA fragmentation, suggesting that the activation of caspases, i.e. those that are inhibited by Z-Asp-CH(2)-DCB, is implicated in the opening of the GSH transporter.     

Bacterial protein toxins acting on intracellular targets

A number of bacterial toxins act on targets located in the cytosol. Diphtheria toxin, Pseudomonas aeruginosa exotoxin A and shigella toxin inhibit protein synthesis by enzymatic inactivation of elongation factor 2 or the 60 S ribosomal subunit. These toxins enter the cells by receptor-mediated endocytosis, followed by translocation across the membranes of intracellular organelles. Also a number or toxins that are not cytocidal act on targets in the cytosol. A number of nontoxic bacterial proteins are able to modify enzymatically intracellular molecules. Some of these proteins could be considered for targeting to special cells followed by translocation to obtain defined physiological effects.     

Pseudomonas exotoxin: chimeric toxins

Pseudomonas exotoxin binds to and enters cells by receptor-mediated endocytosis. Within the cell it requires exposure to low pH to enable it to translocate to the cell cytoplasm where it inhibits protein synthesis by ADP-ribosylating elongation factor 2. The toxin has three main structural domains whose functions are: Ia, cell binding; II, translocation; and III, ADP-ribosylation. Key amino acids have been identified within each domain that are required for the function of the toxin. Chimeric toxins were made originally by using chemical cross-linking reagents to couple Pseudomonas exotoxin (or other toxins) to cell-binding proteins. More recently, a variety of Pseudomonas exotoxin-related chimeric toxins have been made by gene fusion technology. These chimeric toxins may be useful clinically for treating various diseases and experimentally for understanding receptor function.     

Reduction of ricin and other plant toxins by thiol:protein disulfide oxidoreductases

The inhibitory effect of ricin, abrin, and modeccin on protein synthesis by a rabbit reticulocyte lysate is enhanced after preincubation of the toxins with GSH in the presence of a thiol:protein disulfide oxidoreductase purified from bovine liver. The same toxins, as well as the toxin from Viscum album, are reduced also by another thiol:protein disulfide oxidoreductase purified from rat liver cytosol.