Role of the disulfide bond in Shiga toxin A-chain for toxin entry into cells.

Shiga toxin consists of an enzymatically active A-chain and a pentameric binding subunit. The A-chain has a trypsin-sensitive region, and upon cleavage two disulfide bonded fragments, A1 and A2, are generated. To study the role of the disulfide bond, it was eliminated by mutating cysteine 242 to serine. In T47D cells this mutated toxin was more toxic than wild type toxin after a short incubation, whereas after longer incubation times wild type toxin was most toxic. Cells cleaved not only wild type but also mutated A-chain into A1 and A2 fragments. The mutated A-chain was more sensitive than wild type toxin to Pronase, and it was degraded at a higher rate in T47D cells. Subcellular fractionation demonstrated transport of both wild type and mutated toxin to the Golgi apparatus. Brefeldin A, which disrupts the Golgi apparatus, protected not only against Shiga toxin but also against the mutated toxin, indicating involvement of the Golgi apparatus. After prebinding of Shiga(C242S) toxin to wells coated with the Shiga toxin receptor, Gb3, trypsin treatment induced dissociation of A1 from the toxin-receptor complex demonstrating that in addition to stabilizing the A-chain, the disulfide bond prevents dissociation of the A1 fragment from the toxin-receptor complex.

Shiga toxin is produced by Shigella dysenteriae type 1 and consists of an enzymatically active A-chain in noncovalent association with a pentamer of B-chains responsible for binding to Gb 3 , the receptor for Shiga toxin (1)(2)(3)(4)(5)(6). After binding, Shiga toxin is endocytosed and transported to the trans-Golgi network (TGN) 1 (6,7). In sensitive cells the toxin is further transported to the endoplasmic reticulum (8 -10) where translocation of the A-chain to the cytosol may take place. The A-chain is an N-glycosidase, and after entering the cytosol, it cleaves off a single adenine residue from 28 S rRNA of the 60 S ribosomal subunit, resulting in inhibition of protein synthesis (11,12).
The A-chain contains two cysteines (Cys-242 and Cys-261) that form a disulfide bond. The resulting loop can be cleaved by trypsin or furin, separating the A-chain into the enzymatically active A 1 (27.5 kDa) fragment and the A 2 (4.5 kDa) fragment (2,13,14). In cells furin was shown to cleave and thereby activate endocytosed Shiga toxin (14). Cleavage and reduction of Shiga toxin A fragment increase the enzymatic activity, indicating that the presence of the A 2 fragment has an inhibitory effect on the enzymatic activity of Shiga toxin (13). The crystal structure of Shiga toxin has been resolved at 2.5-Å resolution (15), and interestingly, it shows that the active site in A 1 is physically blocked by A 2 . The possibility existed that A 1 and A 2 would form a complex and stay together, even without the disulfide bond. To investigate the role of the disulfide bond in the Achain for intoxication of cells, we have introduced a mutation, C242S, and studied the effect of this A-chain, which does not form a disulfide bond, on different cell lines.
In the present article we demonstrate that the disulfide bond in the A-chain of Shiga toxin is not required for rapid intoxication of cells but seems to be important for toxicity after a long incubation time. The data also suggest that the disulfide bond inhibits dissociation of the A 1 fragment from the toxin complex after cleavage and that it inhibits complete degradation of the A-chain both in vitro and in vivo.

EXPERMENTAL PROCEDURES
Materials-Brefeldin A (BFA) was purchased from Epicentre Technologies (Madison, WI). Pronase, trypsin, Hepes, Gb 3 , N-ethylmaleimide, and PMSF were obtained from Sigma. Na 125 I was purchased from DuPont (Belgium), and [ 3 H]leucine was purchased from Amersham Int. (Amersham, United Kingdom). Furin was a gift from Dr. Gary Thomas (Vollum Institute, Oregon Health Sciences University, Portland, OR). Shiga toxin was purified as described previously (13) and 125 I-labeled as described by Fraker and Speck (16).
Bacteria and Plasmids-Escherichia coli JM109, BMH71-18 mut.S were used for mutagenesis, transformation, recloning into pUC 19, and for isolation of the Shiga toxin. An EcoRI-BamHI restriction fragment if 1.8 kilobase pairs containing ShtA and ShtB was excised from M13 mp8 and subcloned into pALTER (Promega). After mutagenesis, the fragment containing the genes for A and B subunits and the desired mutation in the A subunit was cloned into pUC 19 vector. All bacterial strains were propagated in Luria broth or Luria broth agar, supplemented with antibiotics when necessary. Restriction endonucleases, T4 DNA ligase, and T4 DNA kinase were purchased from New England Biolabs, Inc. (Beverly, MA). Plasmid DNA was purified by methods described by Maniatis et al. (17).
Site-directed Mutagnesis-The point mutation was introduced in the shtA gene by means of the Altered Sites System (Promega). Oligonucleotides designed to introduce mutations were synthesized on the basis of the sequence of the sht as previously published by Kozlov et al. (18). All mutations were confirmed by DNA sequence (19). The oligonucleotides were synthesized on an Applied Biosystem synthesizer and purified with the high pressure liquid chromatography technique. Sequencing primer was 5Ј-GTAAAACGACGGCCAGT-3Ј; the mutagenesis primer was 5Ј-CAACTCGCGATGCATGATGATGAGAATTCAG-3Ј (C242S).
Assay of Shiga Toxin Cytotoxicity-Cells were transferred to 24-wells plates at a density of 3 ϫ 10 4 cells/well 2 or 3 days prior to the experiments. The cells were incubated with increasing amounts of toxin for the indicated time. Then the cells were incubated for 10 min in Hepes medium with 1 Ci/ml [ 3 H]leucine and no unlabeled leucine. The medium was removed, and the cells were washed 2 times in 5% trichloroacetic acid. Finally the cells were dissolved in 0.1 M KOH, and the radioactivity was measured.
Cleavage of Shiga Toxin by Cultured Cells-Cells were seeded out with a density of 6 ϫ 10 4 cells/well in 12-well plates 3 days prior to the experiments. The cells were washed in Hepes medium, and then 125 Ilabeled toxin was added (100 ng/ml) in Hepes medium at 37°C, and the incubation was continued for the indicated period. In other experiments 125 I-labeled toxin was prebound at 0°C for 20 min, the cells were washed in Hepes medium, and the incubation was continued for the indicated period. Then the cells were washed three times with phosphate-buffered saline and lysed in 1% Triton (1% Triton X-100, 20 mM Hepes, 140 mM NaCl, 1 mM phenylmethanesulfonyl fluoride, pH 7.4) on ice for 20 min. The cell lysate was transferred to Eppendorf tubes, nuclei were removed by centrifugation, and proteins were precipitated for 30 min on ice in the presence of 5% trichloroacetic acid. After centrifugation the pellet was washed in ether, dissolved in sample buffer with or without 6% (v/v) 2-mercaptoethanol, and then subjected to SDS-PAGE.
In Vitro Cleavage of Shiga Toxin and Shiga(C242S) Toxin by Purified Furin-The cleavage was performed in a reaction volume of 25 l containing 5 mM CaCl 2 , 1 mM 2-mercaptoethanol, 100 mM buffer (sodium acetate, pH 5.0; MES, pH 5.5-7.5), 10 ng of 125 I-Shiga toxin, and 3 ng of purified furin. The reaction mixture was incubated for 3 h at 37°C, and the reaction was stopped by adding SDS sample buffer with 2-mercaptoethanol. The samples were boiled and subjected to SDS-PAGE.
Trypsin Cleavage of Wild Type Shiga Toxin and Shiga-(C242S) Toxin Bound to Gb 3 -Wells in a 24-well plate were coated with Gb 3 at a concentration of 200 g/ml in 100 l of Hepes medium for 30 min at 37°C and then washed 3 times with Hepes medium, before 100 l of Hepes medium containing 600 ng/ml 125 I-labeled toxin was added. The toxin was allowed to bind for 30 min at room temperature, and unbound toxin was removed by washing 3 times with Hepes medium. Then 100 l of Hepes medium at different pH values (pH 5.5, 6.0, 6.5, 7.0) containing 10 g/ml trypsin was added to each well and incubated for 2 min at room temperature. The supernatant was transferred to an Eppendorf tube on ice containing a final concentration of 5 mM PMSF and sample buffer; the wells were placed on ice and washed once with Hepes medium, and sample buffer with 2-mercaptoethanol and 5 mM PMSF was added. The wells were scraped with a plastic pipette, and the contents were transferred to Eppendorf tubes. The samples were boiled and subjected to SDS-PAGE and autoradiography.
Trypsin Sensitivity of Wild Type Shiga Toxin and Shiga-(C242S) Toxin-125 I-Labeled toxin at a concentration of 800 ng/ml in Hepes medium was treated with increasing concentrations of trypsin (0, 0.2, 1, 5, and 10 g/ml) for 2 min at room temperature. The reaction was stopped by addition of sample buffer containing 2-mercaptoethanol and a final concentration of 5 mM PMSF. The samples were immediately boiled and then subjected to SDS-PAGE and autoradiography.
Pronase Sensitivity of Wild Type Shiga Toxin and Shiga-(C242S) Toxin-125 I-Labeled toxin at a concentration of 800 ng/ml in a reaction volume of 25 l of Hepes medium was incubated with increasing concentration of Pronase (0, 1, 3, 10, 30, and 100 g/ml) for 10 min at room temperature. The reaction was stopped by adding sample buffer containing 2-mercaptoethanol and a final concentration of 5 mM PMSF. The samples were immediately boiled and then subjected to SDS-PAGE and autoradiography.
Acrylamide Gel Electrophoresis-Electrophoresis was carried out as described by Laemmli (21). After electrophoresis the gels were fixed for 30 min in 4% acetic acid and 27% methanol. For autoradiography, Kodak XAR films were exposed to dried gels at Ϫ80°C. After autoradiography, the bands corresponding to A and A 1 of Shiga toxin were quantified by densitometry (model 300A, Molecular Dynamics).
Subcellular Fractionation-The cells were fractionated essentially as earlier described (6). Proteins in fractions corresponding to the Golgi zone, load zone, and endosomal/lysosomal zone were precipitated with 5% trichloroacetic acid, dissolved in sample buffer containing 2-mercaptoethanol, and analyzed by SDS-PAGE and autoradiography.

Toxicity of a Mutated Shiga Toxin (C242S) Lacking the Di-
sulfide Bond in the A-chain-Shiga toxin A-chain contains 2 cysteine residues, cysteine 242 and cysteine 261, linked by an internal disulfide bond. The polypeptide loop between the 2 cysteines contains the sequence Arg-Val-Ala-Arg which is recognized and nicked by trypsin or furin (13,14), thereby separating the A-chain into A 1 and A 2 fragments. To study the role of the disulfide bond in toxin entry into cells, we constructed a mutated Shiga toxin referred to as Shiga(C242S) toxin where cysteine 242 has been mutated to a serine ( Fig. 1). This mutated toxin can no longer form a disulfide bond. To test whether the disulfide bond is required for intoxication of cells, we measured the ability of Shiga(C242S) toxin to intoxicate T47D cells, which are highly sensitive to Shiga toxin. Interestingly, the result showed that Shiga(C242S) toxin was 5-10 times more toxic than wild type toxin when protein synthesis was measured after a 50-min incubation period with toxin continuously present in the medium ( Fig. 2A), whereas after 3 h, wild type toxin was 5-10 times more toxic than the mutated toxin (Fig.  2B). After incubation with toxin for 5 h or overnight the difference was almost 100-fold (data not shown). In HEp-2 cells, wild type toxin and Shiga(C242S) toxin were equally toxic after 50 min incubation, whereas after 3 h incubation wild type toxin exhibited a 20 -30 times higher toxic activity than the mutated toxin (data not shown). To test whether cleavage by furin was responsible for the comparatively low toxicity of the mutated toxin after long incubation times, we tested the effect of the toxins on LoVo cells, which do not contain functional furin (22). Also in LoVo cells wild type toxin was around 30 times more toxic than the mutated toxin after 3 h incubation with cells, and the difference was 100-fold when cells were exposed to toxins overnight (data not shown).
The lack of increase in toxic effect of the mutated toxin after long incubation times with cells could be due to inactivation of the toxin molecules by serum proteases in the medium. To test this possibility wild type toxin and Shiga(C242S) toxin were incubated in serum-containing medium (without cells) overnight at 37°C, and then the ability of the toxins to reduce protein synthesis in a short time experiment was measured. The data showed that overnight exposure to medium with serum did not reduce the toxic effect of the mutated molecule (data not shown). To further investigate the reduced toxicity of the mutated toxin after long incubation times with cells, toxin was prebound to the cells at 0°C, then unbound toxin was washed away, and the incubation was continued for 1.5, 5.5, or 18 h. The cytotoxicity of the wild type toxin increased up to 5 h, whereas the toxic effect of the mutated toxin did not increase after 90 min (Fig. 3). Similar results were also obtained on HEp-2 cells (data not shown). These results suggest that the cells inactivate the mutated toxin more rapidly then the wild type toxin.
Proteolytic Processing of the Wild Type Toxin and Shiga-(C242S) Toxin-Since Shiga(C242S) toxin was less toxic than wild type toxin after long incubation times with cells ( Fig. 3), we tested whether Shiga(C242S) toxin was processed differently than wild type toxin. 125 I-Labeled toxin was prebound to T47D cells at 0°C; then the cells were washed, and the incubation was continued at 37°C for the indicated period of time and analyzed by SDS-PAGE under nonreducing (Fig. 4) and reducing conditions (data not shown), similar to the experiment in Fig. 3. The result showed that both wild type and mutated Shiga toxin A-chain were processed efficiently to A 1 and A 2 fragments. Densitometric measurements of the full-length A and A 1 bands after a 1-h incubation showed that 55% of the wild type A-chain and 47% of the mutated A-chain was processed to A 1 and A 2 fragments. The degradation rate of the mutated A-chain (Fig. 4, lanes 6 and 7) was higher than that of wild type toxin A-chain (lanes 3 and 4). Densitometric measurements of the A and A 1 bands showed that approximately 70% of the cell-associated mutated A-chain was degraded between 1 and 5 h incubation (lanes 6 and 7). In contrast, there was no significant cellular degradation of wild type A-chain in the same period (lanes 3 and 4).
The difference in the rates of degradation could be due to different sensitivity of the molecules to proteolytic enzymes and possibly also to different intracellular sorting. We therefore studied the sensitivity of the wild type and the mutated toxin to trypsin and Pronase treatment. Wild type and mutated A-chain were equally sensitive to cleavage by trypsin which under the condition used induced processing to A 1 and A 2 fragments (Fig. 5). Similarly, processing of the wild type toxin and the mutated toxin to the A 1 fragment occurred at the same concentration of Pronase. After treatment with 1 g/ml Pronase, 20% of the wild type and 22% of the mutated toxin were processed to A 1 (Fig. 6, lanes 2 and 8), and after treatment with 3 g/ml Pronase 53% of wild type and 48% of the mutated toxin were processed (lanes 3 and 9). In contrast, wild type Shiga toxin was more resistant to complete degradation by Pronase than the mutated toxin. In the Pronase sensitivity experiment (Fig. 6), the A and A 1 bands were quantified by densitometry. The result showed that the intensity of the A bands of wild type toxin decreased to 55% when the toxin molecules were treated with 100 g/ml Pronase for 10 min at room temperature. At lower concentrations of Pronase there was no degradation of the wild type A-chain, only processing to form the A 1 fragment. In contrast, the intensity of the A bands of the mutated toxin decreased to 64% at 10 g/ml Pronase. At higher concentra-tions of Pronase the A and A 1 fragments of Shiga(C242S) toxin were almost complete degraded. This shows that Shiga(C242S) toxin is approximately 10 times more sensitive to Pronase treatment than wild type toxin. These results strongly suggest that the disulfide bond protects the A-chain from degradation both in vitro and in vivo.
We have shown earlier that a soluble form of furin cleaves Shiga toxin A-chain with a pH optimum 5.5-6.0, generating A 1 and A 2 fragments (14). To examine whether Shiga(C242S) toxin was cleaved by furin as efficiently and under the same condition as the wild type toxin, we compared cleavage in vitro at different pH values. As shown in Fig. 7, both wild type toxin and the mutated toxin were cleaved to the same extent with a pH optimum at 5.5.
Furin appears to cleave endocytosed wild type toxin most likely in the TGN and/or in endosomes (13,14). We therefore studied the processing of Shiga(C242S) in LoVo cells, both in a stable transfected cell line expressing furin (LoVo/fur) and in a control cell transfected with the vector alone (LoVo/neo). The cells were incubated with 125 I-labeled toxin, and toxin processing was then analyzed by SDS-PAGE and autoradiography. In LoVo/fur cells both wild type A-chain and the mutated A-chain were cleaved efficiently into A 1 and A 2 fragments, and after a 4-h incubation most of the cell-associated toxin molecules were cleaved (Fig. 8b, lanes 3 and 6). In LoVo/neo cells wild type Shiga toxin was cleaved inefficiently, but as shown earlier (14), after long incubation times some cleavage did occur (Fig. 8a,  lane 3). Interestingly, there was essentially no processing of Shiga(C242S) toxin (Fig. 8a, lanes 5 and 6).
To examine the role of the disulfide bond for the integrity of Shiga toxin after cleavage, we used 125 I-labeled toxin bound to wells coated with Gb 3 , the receptor for Shiga toxin. The toxin did not bind to wells not coated with Gb 3 (data not shown). After binding, unbound toxin was washed away, and trypsin (10 g/ml in Hepes medium) was added for 2 min at room temperature to cleave the toxin. Then the supernatant was removed, and SDS sample buffer was added to the well to solubilize the receptor-bound toxin molecules. Proteins both in the supernatant and bound to the plastic were subjected to SDS-PAGE and autoradiography. The experiment was performed at pH 5.5-7.0 to see if pH affected the release of A 1 after cleavage. The low pH was used to mimic the low pH in endosomes or TGN. The results showed that upon trypsin treatment the A 1 fragment of Shiga(C242S) toxin was released from the toxin-receptor complex after cleavage at all pH values tested (Fig. 9A, lanes 4, 6, 8, and 10). In contrast, when a similar experiment was performed with wild type toxin, the A 1 fragment was not released (Fig. 9B, lane 4, 6, 8, and 10). In all cases the B-chains were still attached to the wells with Gb 3 (Fig. 9, A  and B, lanes 3, 5, 7, and 9). As expected, Shiga(C242S) toxin precleaved with trypsin had reduced cytotoxicity to T47D cells (data not shown).
Intracellular Trafficking of the Toxin Molecules-It has been shown previously that Shiga toxin is transported to the Golgi apparatus (7,23), and BFA, which disrupts the Golgi apparatus in several cell lines (23)(24)(25)(26)(27)(28)(29)(30)(31)(32), protects cells against Shiga toxin (13). This suggests that transport to the Golgi apparatus is required for intoxication. To investigate whether Shiga-(C242S) toxin follows the same endocytic pathway to the cytosol, we tested the effect of BFA also on intoxication with this molecule. We found that BFA also protected against Shiga-(C242S) toxin (Fig. 10). Furthermore, subcellular fractionation of cells incubated with 125 I-labeled toxin showed that approximately 6% of both wild type toxin and Shiga(C242S) toxin were transported to the Golgi apparatus during 1 h (data not shown). Importantly, as shown in Fig. 11, uncleaved Shiga(C242S) toxin A-chain reaches the Golgi apparatus (lane 1). DISCUSSION We have introduced a mutation in Shiga toxin A-chain, C242S, to eliminate the disulfide bond in the A-chain, to study the role of the disulfide bond for toxin entry into cells, and for stabilization of the A-chain. Our data show that the disulfide bond is not strictly required for intoxication of cells but is important for efficient intoxication of cells after long incubation times. The results also indicate that the disulfide bond inhibits degradation of the A-chain and inhibits dissociation of the A 1 fragment from the toxin-receptor complex after cleavage.
Shiga toxin A-chain is easily cleaved by furin both in vitro and in vivo separating the A-chain into A 1 and A 2 fragments connected by a disulfide bond (14). In Shiga toxin furin recognizes the sequence Arg-Val-Ala-Arg (14), which is a recognition motif for the membrane-anchored protease furin (33). However, cleavage by furin might also be dependent on protein conformation, and it was not clear whether the enzyme would cleave the mutated molecule under the same conditions as the wild type toxin. As shown here, a soluble form of furin cleaved both the mutated and wild type toxin with the same pH optimum. Furthermore, Shiga(C242S) toxin A-chain was also efficiently cleaved in LoVo cells transfected with furin, whereas in control cells, which do not express functional furin (22), the cleavage was strongly reduced. This demonstrates that furin is able to cleave the mutated toxin also in vivo. Earlier studies have shown that in LoVo cells not transfected with furin (14) wild type toxin is cleaved slowly into A 1 and A 2 fragments and that an inhibitor of the cytosolic protease calpain both inhibited cleavage and protected LoVo cells against Shiga toxin. Therefore, a protease other than furin, possibly calpain, may cleave and activate Shiga toxin in cells with low furin activity. Calpain has been reported to recognize several different amino acid sequences at the cleavage site, suggesting that other characteristics such as amino acids at a distance from the cleavage site or the protein conformation may be important (34,35). Thus, the reduced cleavage of the mutated toxin compared with wild type toxin in LoVo cells could be due to lack of recognition of the mutated toxin by calpain.
The disulfide bond in the A-chain of Shiga toxin is not strictly required for intoxication of cells. Interestingly, Shiga (C242S) toxin intoxicated T47D cells more efficiently than wild type toxin after short incubation times. However, after longer incubation times wild type toxin was more toxic than the mutated toxin. These surprising results could be due to different intracellular sorting of the two toxins. However, brefeldin A protected T47D cells not only against wild type toxin (13) but also against the mutated toxin in a short incubation time with cells. Furthermore, subcellular fractionation of cells showed that both wild type and mutated toxin were transported to the Golgi apparatus to the same extent. The results therefore indicate that the pathway used for intoxication of the cells is the same for the two toxins.
Cell-associated Shiga(C242S) toxin clearly intoxicates cells FIG. 9. Dissociation of A 1 fragment from the toxin-receptor complex after cleavage of Shiga(C242S) toxin with trypsin. Wells in a 24-well plate were coated with Gb 3 (200 g/ml) in Hepes medium for 30 min at 37°C, and then the wells were washed with Hepes medium to remove unbound Gb 3 , and 125 I-labeled toxin (600 ng/ml) in Hepes medium was added at room temperature in the wells and incubated for 30 min. Unbound toxin was washed away with Hepes medium, and 100 l of Hepes medium with trypsin (10 g/ml) at pH 5.5, 6.0, 6.5, 7.0 was added per well and incubated for 2 min at room temperature. The supernatant (S) was removed and sample buffer with 2-mercaptoethanol and a final concentration of 5 mM PMSF was added to the supernatant. Proteins still bound to Gb 3 on the plastic (P) were solubilized with sample buffer with 5 mM PMSF. Both proteins in the supernatant and bound to Gb 3 were analyzed by SDS-PAGE (12%). less efficiently than wild type toxin in long incubation times. This is not due to inactivation of Shiga (C242S) toxin by serum proteases in the incubation medium, but seems to be due to cell-mediated inactivation only. In fact, studies in cells showed that the degradation rate of the mutated A-chain was higher than that of the wild type A-chain. Also, the mutated A-chain was more sensitive to Pronase treatment than wild type Achain. These data suggest that the normal resistance of the wild type toxin to proteases (2) is dependent on the formation of the disulfide bond of the A-chain. Thus, the reduced toxicity of the mutated toxin compared with wild type toxin after long incubation times can be explained by increased cellular degradation of the mutated toxin. We can only speculate on the reason for the strong toxic effect of the mutated toxin on T47D cells after short incubation times. It is possible that the slow reduction rate of the disulfide bond that occurs in the A-chain of the wild type toxin (Ref. 13, also shown in Fig. 4) will delay the action of the toxin sufficiently so that the mutated toxin despite the higher degradation rate can be more toxic than wild type toxin in a short incubation time with cells. This might especially occur in cells like T47D cells where retrograde transport of Shiga toxin to the endoplasmic reticulum may be especially efficient (13).
It appears from the crystal structure of Shiga toxin that the active site cleft in A 1 is blocked by a segment of the A 2 fragment (15). However, as shown here, possible interactions between the A 1 and A 2 fragments of Shiga toxin are not sufficiently strong to inhibit the dissociation of the A 1 fragment from the toxin-receptor complex of Shiga(C242S) toxin. When the toxin molecule was prebound to a well coated with the receptor for Shiga toxin, cleavage induced rapid dissociation of the A 1 fragment from the toxin-receptor complex. In accordance with this, trypsin-nicked mutated toxin was less toxic to T47D cells than unnicked toxin. In contrast to Shiga toxin, the A 1 fragment of cholera toxin does not seem to dissociate easily from the toxin complex after reduction of the A-chain (36).
Shiga(C242S) toxin is, similar to the wild type toxin (14), most likely cleaved in the TGN and/or in endosomes, and the A 1 fragment may dissociate from the toxin-receptor complex at this location and therefore not intoxicate cells efficiently. Subcellular fractionation of cells showed that uncleaved Shiga(C242S) toxin is able to reach the Golgi apparatus, suggesting that the A-chain can be further transported to the location in the cells where it is translocated to the cytosol, most likely in the endoplasmic reticulum (8 -10). Since the mutated toxin is also less toxic than wild type toxin in LoVo cells, which do not process the mutated A-chain into A 1 and A 2 fragments, the data suggest that increased degradation is more important than furin-induced dissociation of the A 1 -chain for the decreased ability to intoxicate cells. The data support the view that the disulfide bond has an important role in keeping the A-chain assembled after cleavage and that it protects the Afragment from degradation.