Studies on protein uptake by isolated tumor cells. 3. Apparent stimulations due to pH, hypertonicity, polycations, or dehydration and their relation to the enhanced penetration of infectious nucleic acids.

Fixation of (131)I-serum albumin by Ehrlich ascites tumor cells in suspensions and sarcoma S-180 monolayers was measured under experimental conditions. Anaerobic incubation and inhibitors of the oxidative metabolism critically restricted the range of glucose concentrations capable of supporting cell life; in glucose concentrations higher than 10(-2)M, Ehrlich cells suffered from their own acid production; in concentrations 10(-2)M, lower than they underwent damage by starvation. Both types of damage were accompanied by increased albumin fixation unrelated to pinocytosis. Different procedures recommended to enhance the uptake of infectious viral RNA by animal cells in culture were tested for their ability to increase albumin uptake. They enhanced the penetration of both albumin and vital dyes and decreased the viability of cell populations. Their effect, therefore, is related to cell damage. It was postulated that reversible damage to cells favors RNA infection by leading to abnormal uptake processes and by decreasing intracellular digestion. This abnormal uptake is different from pinocytosis and also from the massive fixation of albumin to dead cells. The latter phenomenon is due to adsorption by intracellular sites exposed by disruption of the cell membrane. Polycations are able to induce all three forms of fixation depending on the experimental conditions.

Uptake of protein by mammalian cells: an underdeveloped area. The penetration of foreign proteins into mammalian cells can be measured and their functions explored.

Although it is accepted on the basis of biological and morphological evidence that mammalian cells will take up macromolecules, little is known about the kinetics, the specificity, and the functions of this uptake. With labeled proteins used as models, it is found that the transport proceeds at very low rates, requires little energy, and is markedly enhanced by polybasic compounds. Molecular charge and size are important factors: cells clearly favor cationic macromolecules of large molecular weights. Neither factor, however, can fully account for the selectivity detected in the uptake of different proteins. Ingested albumin undergoes rapid and extensive degradation. This fact suggests that macromolecules have only a limited chance to express their biological activity in target cells, a finding that is relevant also to the role of foreign nucleic acids and the possibility of achieving genetic transformation in animal cells. There are concrete indications, however, that in spite of their short half-life, proteins can act as carriers, as precursors of active agents, and as regulators of metabolic functions in host cells. They may also be important in the control of growth and differentiation. These functions of exogenous proteins are still largely unexplored.

Marked synergism of dimethylnitrosamine carcinogenesis in rats exposed to cadmium.

The combined carcinogenic effect of cadmium and dimethylnitrosamine (DMN) was examined in male Wistar rats, to test whether the previously observed synergism in mutagenicity between cadmium and 2 N-nitroso compounds would be paralleled by a synergism in carcinogenicity. In experiment 1, 50 five-month old rats received 18 mg DMN/kg i.p. followed by cadmium i.m. in two injections totaling either 1.5 or 3.0 mg Cd2+/kg. In experiment 2, 30 weanling rats received five i.m. doses totaling 6 mg Cd2+/kg followed by 18 mg DMN/kg. After 52 weeks, the incidence of renal tubular neoplasms was significantly increased above additivity in both experiments (P = 0.0005 to 0.017). Experiment 1 also showed a synergistic increase in the incidence of neoplastic (P = 0.024) and preneoplastic (P less than 0.01) microscopic liver lesions, of tumors of organs other than liver and kidney (P less than 0.01), of malignant versus benign tumors (P = 0.038), and of multiple versus single tumors (P = 0.0019). In experiment 2, the incidence of DMN-induced hepatocellular adenomas was significantly lower than additivity, suggesting an antagonistic, protective effect of cadmium pretreatment. The overall incidence of tumors of any type was 17.5 versus 67.7% in control (one agent or none) versus test animals. Malignant tumors included carcinomas, sarcomas, and lymphomas, involving nine different sites. Cadmium appears to enhance the initiation of carcinogenesis induced by DMN. Both cadmium and nitrosamines are known environmental contaminants present in air, food, water supplies, and tobacco smoke.

Targeting, internalization, and cytotoxicity of methotrexate-monoclonal anti-stage-specific embryonic antigen-1 antibody conjugates in cultured F-9 teratocarcinoma cells.

Methotrexate (MTX) conjugates of a monoclonal antibody, anti-SSEA-1, containing an average of 45 mol MTX/mol of immunoglobulin M, were prepared by a carbodiimide coupling reaction. Binding experiments indicate that conjugation does not decrease the affinity of the antibody for its antigen. The conjugate strongly inhibits the growth of SSEA-1-bearing F-9 teratocarcinoma cells, with 50% inhibitory dose of 4.5 nM MTX, which makes it more active than free MTX (50% inhibitory dose of 15 nM). The drug-free antibody is not cytotoxic to F-9 cells at the concentrations used. The high efficacy of the conjugated drug may be due in part to the fact that anti-SSEA-1 antibody is an immunoglobulin M. MTX conjugated to nonspecific immunoglobulin M has little inhibitory effect (50% inhibitory dose of 150 nM). When acting on SSEA-1 negative cells, the two conjugates have only a small but identical effect. Thiamine pyrophosphate, an inhibitor of MTX transport, can prevent the cytotoxicity of the free MTX but not that of the anti-SSEA-1 conjugate. Leupeptin, an inhibitor of lysosomal protease, can partially protect F-9 cells against the antibody conjugate but not against free MTX. These results indicate that the MTX antibody conjugate binds specifically to F-9 cells, and is internalized and intracellularly degraded to release a small molecular active drug. Pretreatments of F-9 cells for 1 h with unlabeled antibody inhibits the subsequent uptake of identical concentration of labeled conjugate. The rate of internalization, however, regains almost normal values within 4 h, indicating a rapid reappearance of free antigenic sites at the cell surface.

Binding sites and endocytosis of heparin and polylysine are changed when the two molecules are given as a complex to Chinese hamster ovary cells.

This paper characterizes the complex formed in vitro between polylysine and heparin in the presence of heparin excess, and investigates the interaction of this complex with the surface of Chinese hamster ovary cells. It examines the kinetics of surface binding and cellular uptake of the complex and shows that both processes can be distinguished from those of free heparin and free polylysine. The view that these three ligands bind to different surface sites is further supported by the fact that their interaction with cells is influenced differently by cell detachment with trypsin, detachment with EGTA or exposure to acid pH. Membrane transport of the complex is a saturable process suggestive of receptor-mediated endocytosis. It is, however, less effective than would be expected on the basis of the binding kinetics. Only 40% of the complex bound at 0 degrees C is internalized during a 2 h reincubation period at 37 degrees C, suggesting some degree of uncoupling between binding and endocytosis. These data confirm prior results obtained with methotrexate-polylysine conjugates. We had shown that the addition of heparin to a medium containing a methotrexate-polylysine conjugate leads unexpectedly to a marked cellular uptake of drug conjugate, which is capable of killing cells that are otherwise resistant to free methotrexate (Shen, W.-C. and Ryser, H.J.-P. (1981) Proc. Natl. Acad. Sci. USA 78, 7589-7593). The polylysine X heparin complex is therefore of interest as a potential carrier for intracellular drug delivery through endocytosis.

Mutagenicity of cadmium in Salmonella typhimurium and its synergism with two nitrosamines.

Cadmium chloride (CdCl2) at concentrations of 0.5 mM was significantly mutagenic in Salmonella typhimurium tester strains and reverted histidine auxotrophy due either to missense (TA1975 and TA1535) or to frameshift (TA1537) mutations. It also induced forward mutations to 8-azaguanine resistance in each strain, but failed to increase mutation frequencies in strain TA100. More importantly, CdCl2 increased the mutagenicity of two common nitrosamines in synergistic fashion, at a level up to 30-fold greater than expected from simple additivity. The mutation frequency induced by N-methyl-N'-nitro-N-nitrosoguanidine was increased about 10-fold in the presence of 0.5 mM CdCl2. This synergism was seen both in the induction of 8-azaguanine resistance and the reversion of histidine auxotrophy and was observed in the repair-proficient strain TA1975 as well as its repair-defective (uvrB-) derived strain TA1535. The synergism was dependent upon Cd concentration and was much reduced at 0.25 mM CdCl2. The strongest synergism was observed in the reversion of histidine auxotrophy in TA1975 by 180 microM methylnitrosourea and 0.5 mM CdCl2. In contrast to mutagenicity, there was no evidence for synergism in the toxicity of CdCl2. These data suggest that cadmium might interfere with the repair of both spontaneous and nitrosamine-induced mutations. They also raise the possibility that cadmium and nitrosamines may have synergistic effects as environmental carcinogens.

Mechanisms of synergism in the mutagenicity of cadmium and N-methyl-N-nitrosourea in Salmonella typhimurium: the effect of pH.

Cadmium enhances the mutagenic effect of N-methyl-N-nitrosourea (MNU) in a synergistic manner in Salmonella typhimurium. In the range of doses that yield synergistic effects, it is by itself highly cytotoxic and only weakly mutagenic. A decrease in pH from 7 to 6 markedly decreases cadmium toxicity, causing a 4-fold increase in the surviving fraction. The same shift in pH markedly increases the dose-dependent mutagenesis of MNU, whether MNU is acting alone or in combination with cadmium and increases the synergism at lower doses of cadmium. Thus, the synergism appears to depend on the mutagenicity of MNU and not on the cytotoxicity of Cd. The combined mutagenic effect of Cd and MNU is comparable in tester strain TA1535 and TA100, which contains the error-prone (SOS) repair-enhancing pKM101 plasmid. Thus the synergistic effect of cadmium is not enhanced by the induction of SOS processing. The hydrolysis of MNU is higher at pH 7 than at pH 6 but cadmium has no effect on the rate of MNU hydrolysis at either pH, and therefore does not influence the concentration of the active electrophile of MNU. Cadmium does not enhance the mutagenesis of ethylnitrosourea, the ethylating analogue of MNU, indicating that the synergism is specific for methylated DNA lesions. These data suggest that cadmium acts either directly by modifying the nature or indirectly by inhibiting the repair of methylation damage.

Poly(L-lysine) has different membrane transport and drug-carrier properties when complexed with heparin.

Methotrexate (MTX) conjugated to a Mr 3000 poly(L-Lys) markedly inhibits the growth of Pro-3 MtxRII5-3 Chinese hamster ovarian cells, a mutant line known to be drug resistant because of defective MTX transport. In these cells, membrane transport of [3H]MTX-poly(Lys) is sharply decreased by addition of 0.5- to 2.5-fold heparin but remains at 15-20% of control in 2.5- to 50-fold heparin excess. Heparin addition at first markedly inhibits but, at high concentration, restores the growth inhibitory effect of MTX-poly(Lys). In excess heparin, MTX-poly(Lys) is transported as a heparin complex. Because reduced transport (15-20%) is sufficient to cause a 90% inhibition of cell growth, MTX-poly(Lys) apparently gains pharmacologic potency when compared to heparin. This gain can be related to a greater inhibitory effect on dihydrofolate reductase and to a different mode of transport. The inhibitory effect of MTX-poly(Lys) on dihydrofolate reductase in vitro is increased nearly 100-fold in the presence of excess heparin but remains less than that of free MTX. Unlike that of MTX-poly(Lys), the transport of MTX-poly(Lys)-heparin has the characteristics and efficiency of a receptor-mediated process. It proceeds by endocytosis but is not, as in the case of uncomplexed conjugate, followed by the intracellular generation of pharmacologically active breakdown products that would account for cytotoxicity. These observations raise the possibility that at least part of the MTX-poly(Lys)-heparin reaches cellular dihydrofolate reductase in the form of macromolecular complexes that escape from entrapment in endocytotic structures. Our data illustrate a way to overcome drug resistance by taking advantage of the specific uptake of a macromolecular drug carrier. They offer a method of drug delivery in which heparin improves selectivity and decreases the unwanted toxicity inherent to polycationic carriers.

Selective killing of Fc-receptor-bearing tumor cells through endocytosis of a drug-carrying immune complex.

A soluble immune complex was used as a drug carrier targeted to Fc-receptor-positive cells. Two receptor-positive tumor cell lines, WEHI-3 and M5076, were exposed to methotrexate-human serum albumin conjugate (MTX-HSA) in the presence and absence of anti-HSA antiserum. Both cell types were killed by 30 nM MTX when the drug conjugate was given in the presence of antiserum but were totally unaffected in the absence of antiserum. Drug-free HSA given with antiserum had no effect. Both cell lines responded similarly despite their marked difference in phagocytotic activity. One of the two lines, M5076, is defective in MTX transport and hence resistant to free MTX. Since this line would not be affected by MTX released extracellularly from MTX-HSA, its susceptibility implies that MTX is released inside cells, after endocytosis of the complex, and that endocytosis circumvents the transport defect. Two cell lines lacking Fc receptors (CHO and L929) were not influenced by the drug complex. The pharmacologic effect is mediated by a specific ligand-receptor interaction, since Fc receptor-positive cells are protected by an excess of unconjugated HSA and by the addition of a small amount of staphylococcal protein A, which binds to the Fc portion of IgG. These data demonstrate that Fc receptors can be exploited for cellular drug delivery using a common antigen-antibody complex as a drug carrier.

Conjugation of methotrexate to poly (L-lysine) as a potential way to overcome drug resistance.

Methotrexate (MTX) was conjugated through a carbodiimide-catalyzed reaction to poly (L-lysines) of various molecular sizes at a ratio of approximately one molecule per 27 lysyl residues. These conjugates were tested on cultured, Chinese hamster, ovary cells known to be drug resistant because of a deficient methotrexate transport. The cellular uptake of conjugated drug far exceeded the uptake of free drug in both drug sensitive and resistant lines. The conjugated drug inhibited the growth of the transport-deficient cells at concentrations at which free drug had no effect. The conjugate failed to inhibit dihydrofolate reductase in vitro. This and other evidence indicate that the strong pharmacologic effect of the MTX-poly(Lys) conjugate is due to the intracellular--presumably intralysosomal--hydrolysis of its polymeric backbone followed by the release inside the cell of an active form of MTX. This conclusion is supported by data obtained with conjugates using poly(D-lysine) as a carrier. This optical isomer is not susceptible to common proteolytic enzymes and MTX-poly(D-Lys) has not growth inhibitory effect whatever on either transport proficient or deficient CHO-cells. MTX-poly(L-lys) can thus be seen as a lysosome-activated drug.

Conjugation of methotrexate to poly(L-lysine) increases drug transport and overcomes drug resistance in cultured cells.

Methotrexate and [(3)H]methotrexate were conjugated through a carbodiimide-catalyzed reaction to a 70,000 molecular weight poly(L-lysine) in molar ratios of approximately 13 to 1. The cellular uptake of labeled conjugate was far in excess of the uptake of free drug in cells that were either proficient or deficient in methotrexate transport. The conjugate markedly inhibited the growth of PRO(-)3 Mtx(RII) 5-3 Chinese hamster ovary cells, which are known to be drug resistant by virtue of a deficient methotrexate transport. The cells, however, were not inhibited by the same concentrations of free poly(Lys) and free drug. The 100-fold difference in drug concentration needed to inhibit the mutant cells and their corresponding wild type was totally abolished by exposing the methotrexate-resistant cells to methotrexate-poly(Lys). That the drug is carried into the resistant cells as intact drug-poly(Lys) is evident also from the fact that the conjugate is rendered inactive by brief trypsinization in vitro. Because the conjugate fails to inhibit dihydrofolate reductase (5,6,7,8-tetrahydrofolate: NADP(+) oxidoreductase; EC 1.5.1.3) in vitro, it must be concluded that the strong growth inhibitory effect of the conjugate is due to the intracellular hydrolysis of its polymeric backbone, followed by the release inside the cell of a pharmacologically active form of methotrexate. Our date show that in methotrexate-resistant cells the intracellular release of active drug after uptake of conjugate is of the same order of magnitude as the uptake of free drug by transport-proficient cells and, hence, that the drug resistance due to deficient transport can be totally overcome.

Conjugation of poly-L-lysine to albumin and horseradish peroxidase: a novel method of enhancing the cellular uptake of proteins.

The carbodiimide-catalyzed conjugation of a 6700 molecular weight fragment of poly-L-lysine to radiolabeled human serum albumin or to horseradish peroxidase enhances the membrane transport of each protein into cultured mouse fibroblasts approximately 11- and 200-fold, respectively. At least 50% of the peroxidase activity remained after conjugation. Trypsinization and carbamylation of the two conjugates demonstrates that the enhancement of their cellular uptake is related to their poly-L-lysine content. Simple addition to the medium of comparable amounts of free poly-L-lysine has no effect on the transport of either native protein. Addition of poly-L-ornithine (molecular weight 200,000) at 3-30 microgram/ml, a condition known to cause enhancement of 125I-labeled human serum albumin uptake by mouse sarcoma cells, has no visible effect on the cellular uptake of native horseradish peroxidase. The intracellular localization of the enzyme-poly-L-lysine conjugate can be demonstrated cytochemically by either light or transmission electron microscopy. A concentration of conjugate that increases the uptake more than 200-fold does not cause any detectable morphological change suggestive of cell toxicity. Furthermore, because poly-L-lysine is an excellent substrate for intracellular proteolytic enzymes, it can be expected to be broken down and reutilized in the cell.

Heparin interaction with cultured cells: possible role of fibronectin in uncoupling surface binding and endocytosis.

Cellular uptake of 3H-heparin by cultured Chinese hamster ovary cells does not follow the pattern that would be expected on the basis of binding kinetics. It is essentially non-saturable and its low pinocytotic index is consistent with fluid-phase endocytosis, while surface binding is saturable and suggests the presence of specific receptors. Most of the surface-bound heparin is released into the medium and only minimal amounts are internalized when temperature is shifted from 0 degree to 37 degrees C. This uncoupling of surface binding and endocytosis may be due to the binding of heparin to an external surface macromolecule that is only loosely attached to the plasma membrane, and may therefore escape endocytosis. Fibronectin is a likely candidate to provide that type of binding site. Heparin has been found to bind specifically to a number of mammalian cells (Kjellen et al., 1977; Glimelius et al., 1978; Shanberge et al., 1976; Fabian et al., 1978), but no attempt was made in those studies to distinguish between surface binding and cellular uptake. Heparin is transported into cells when complexed to other macromolecules (Fabian et al., 1978; Shen and Ryser, 1981; Morad et al., 1982) and does then exert some cellular effects (Shen and Ryser, 1981). Despite the wide clinical use of heparin, little is known about its membrane transport and intracellular effects, although endocytosis of other polyanions has been shown to inhibit phago-lysosomal fusion in macrophages (Geisow et al., 1980).(ABSTRACT TRUNCATED AT 250 WORDS)

Colorimetric determination of microgram quantities of polylysine by trypan blue precipitation.

A simple and sensitive method for the determination of polylysine in solution is described. Polylysine is quantitatively precipitated with trypan blue. The absorbance of unbound dye in the supernatant is inversely proportional to the concentration of this polyamino acid. The precipitation is identical for all sizes of polylysine of molecular weight 13,000 or higher, and is prevented by the addition of either polyanions or serum. The measurable range of polylysine hydrobromide is between 1 and 10 micrograms/ml, which is about 10-fold lower than that by the published methyl orange precipitation method.

The cellular uptake of horseradish peroxidase and its poly(lysine) conjugate by cultured fibroblasts is qualitatively similar despite a 900-fold difference in rate.

When horseradish peroxidase (HRP) is conjugated to poly(L-lysine) of molecular weight (MW) 13,000, its transport into cultured L929 fibroblasts in 1 hour at 37 degrees C is increased 918-fold. The kinetics of uptake are linear with time and concentration, reflecting a process of nonreceptor-mediated adsorptive endocytosis. Neither HRP-poly(Lys) conjugate nor free poly(Lys) of any size cause any increase in fluid phase endocytosis. All evidence indicates that the covalently bound poly(Lys) increases the binding of HRP to the cell surface and that this adequately accounts for an increased internalization occurring at a steady rate. In the absence of Ca++, both surface binding and uptake of the conjugate, but not of free HRP, are decreased. Cytochalasin B does not significantly inhibit the transport of either form of HRP. The half-lives of HRP and HRP-poly(Lys) are 7.6 and 5.6 hours, respectively, when measured over a period of 12 hours. Electron microscopic analysis of cells that have ingested comparable amounts of HRP shows that the intracellular localization of free and conjugated HRP are comparable and that both are found inside the same array of structures. Thus HRP, which binds very poorly to the cell surface and is considered a fluid-phase marker for the "contents" of pinocytotic vesicles, and HRP-poly(Lys), which binds very strongly to the cell surface and is considered a membrane marker for adsorptive endocytosis, are taken up along the same endocytotic pathway. Moreover, despite the fact that neither markers are transported by receptor-mediated endocytosis, both are seen in structures that were described as receptosomes. It appears, therefore, that the pathways utilized in receptor-mediated transport are available to all other forms of pinocytosis.

Disulfide spacer between methotrexate and poly(D-lysine). A probe for exploring the reductive process in endocytosis.

Poly(D-lysine) is taken up avidly by cultured cells through adsorptive endocytosis and can serve as a carrier to increase cellular uptake of other molecules. While direct conjugation of methotrexate to poly(D-lysine) yields a conjugate devoid of cytotoxic effects because poly(D-lysine) is not digested in lysosomes, the indirect conjugation using a triglycine spacer or a disulfide spacer strongly inhibits the growth of both the wild type and the methotrexate transport-defective Chinese hamster ovary cells. Cell treatment with 3 mM NH4Cl or 50 micrograms/ml leupeptin prevents the effect of conjugate with the triglycine spacer, but not of conjugate with the disulfide spacer. On the other hand, preincubation with 2-mercaptoethanol abolishes the effect of the drug-disulfide conjugate in the methotrexate transport-defective mutant, but not the effect of the drug-triglycine conjugate. The disulfide conjugate shows an identical cytotoxic effect in alpha-minimal essential medium and RPMI 1640 media, even though cells grown in the latter have only half the glutathione content as cells grown in the former medium. We conclude that the reductive process through which methotrexate is released from the disulfide spacer (a) occurs inside cells and not at the cell surface, (b) requires neither acid pH nor lysosomal enzymes, and (c) is not mediated by a glutathione-disulfide exchange reaction requiring high glutathione concentrations. Although the cellular compartment in which this reductive process occurs is not yet identified, there are reasons to assume that it is prelysosomal.

Altered endocytosis in a mutant of LM fibroblasts defective in cell-cell fusion.

Mutants of LM fibroblasts selected for their decreased ability to undergo polyethylene glycol-induced cell-to-cell fusion (F40 subline) were examined for possible alterations of their ability to carry out endocytosis. Both fluid-phase endocytosis of inulin and horseradish peroxidase and nonreceptor mediated adsorptive endocytosis of poly(L-lysine) were reduced to 60% of control values. Comparable results were obtained when the uptake of poly(L-lysine) was measured as internalization of surface-bound label in label-free medium or following continuous exposure. Accelerated breakdown of internalized label was ruled out as a cause for decreased label accumulation. Accelerated exocytosis is an unlikely cause, and it is suggested that the decreased uptake is due to a decrease in the constitutive membrane vesiculation process that leads to the formation of endocytotic vesicles. The capacity of F40 cells to degrade internalized horseradish peroxidase and poly(L-lysine) was not impaired, nor was their susceptibility to the cytotoxic action of methotrexate-poly(L-lysine). This drug conjugate must be degraded inside cells and release small molecular methotrexate in order to be cytocidal. These data suggest that only the first step of nonspecific endocytosis is impaired, while the subsequent steps that require fusion of endosomes to lysosomes proceed normally. Since the formation of primary endosomes requires membrane fusion through the external aspect of the plasma membrane and in that respect resembles cell-cell fusion, we propose the hypothesis that the observed decrease in endocytosis is related to the decreased ability of F40 cells to fuse with each other, and reflects a decreased efficiency of fusion processes at the external face of the plasma membrane.

Cell surface sulfhydryls are required for the cytotoxicity of diphtheria toxin but not of ricin in Chinese hamster ovary cells.

A previous study on cleavage of disulfide bonds in endocytosed model compounds had shown that an initial phase of cleavage was totally inhibited by membrane-impermeant sulfhydryl inhibitors and thus was mediated by cell surface sulfhydryls (Feener, E. P., Shen, W.-C., and Ryser, H. J.-P. (1990) J. Biol. Chem. 265, 18780-18785). This paper uses the same inhibitors (5,5'-dithiobis(2-nitrobenzoic acid) and p-chloromercuriphenylsulfonic acid) to examine the role of surface sulfhydryls in the cytotoxicity of diphtheria toxin (DT). Since the interchain disulfide of endocytosed DT must be cleaved prior to translocation of chain A from endosomes to cytoplasm, it was postulated that surface sulfhydryls might mediate the cleavage of that disulfide bond as well. Both sulfhydryl blockers did indeed markedly inhibit DT cytotoxicity. This effect was not due to inactivation of unbound DT, inhibition of receptor-mediated endocytosis, or impairment of acidification of endosomes. We conclude that cell surface sulfhydryls susceptible to blockage by 5,5'-dithiobis(2-nitro-benzoic acid) and p-chloromercuriphenylsulfonic acid are required for the cytotoxicity of DT and, most likely, for the reductive cleavage of DT's interchain disulfides. Ricin cytotoxicity was not decreased; this is consistent with the view that ricin reaches the cytoplasm from a late endocytic structure and with the finding that endocytosed disulfides are also cleaved in a cell fraction containing elements of the Golgi apparatus (Feener, E. P., Shen, W.-C., and Ryser, H. J.-P. (1990) J. Biol. Chem. 265, 18780-18785).