Identification of an Essential Region for Translocation of Clostridium difficile Toxin B

Clostridium difficile toxin A (TcdA) and toxin B (TcdB) are the major virulence factors involved in C. difficile-associated diarrhea and pseudomembranous colitis. TcdA and TcdB both contain at least four distinct domains: the glucosyltransferase domain, cysteine protease domain, receptor binding domain, and translocation domain. Few studies have investigated the translocation domain and its mechanism of action. Recently, it was demonstrated that a segment of 97 amino acids (AA 1756–1852, designated D97) within the translocation domain of TcdB is essential for the in vitro and in vivo toxicity of TcdB. However, the mechanism by which D97 regulates the action of TcdB in host cells and the important amino acids within this region are unknown. In this study, we discovered that a smaller fragment, amino acids 1756–1780, located in the N-terminus of the D97 fragment, is essential for translocation of the effector glucosyltransferase domain into the host cytosol. A sequence of 25AA within D97 is predicted to form an alpha helical structure and is the critical part of D97. The deletion mutant TcdB∆1756–1780 showed similar glucosyltransferase and cysteine protease activity, cellular binding, and pore formation to wild type TcdB, but it failed to induce the glucosylation of Rho GTPase Rac1 of host cells. Moreover, we found that TcdB∆1756–1780 was rapidly degraded in the endosome of target cells, and therefore its intact glucosyltransferase domain was unable to translocate efficiently into host cytosol. Our finding provides an insight into the molecular mechanisms of action of TcdB in the intoxication of host cells.


Introduction
Clostridium difficile is the major cause of antibiotic-associated diarrhea and pseudomembranous colitis worldwide. Two exotoxins, toxin A (TcdA) and toxin B (TcdB), are the major virulence factors involved in C. difficile infection (CDI) [1,2], and both belong to the family of clostridial glucosylating toxins. The toxins are multi-domain proteins containing at least four functional domains [3]. The N-terminus of the toxin harbors the glucosyltransferase domain (GTD) that inactivates host Rho GTPases by glucosylation [4,5] and a cysteine protease domain (CPD) responsible for autoprocessing [6][7][8]. The C-terminus, consisting of combined repetitive oligopeptides (CROP), is predicted to be a receptor binding domain (RBD) [9,10]. The receptor for TcdB has been identified recently [11,12], but additional receptors may exist [13,14]. A large region between the CPD and RBD is thought to be the translocation domain (TD) which is important for delivery of N-terminal enzymatic domains into the host cytosol via pore formation [15][16][17][18][19].

Prediction of the Secondary Structure of the D97 Segment
To gain insight into the structure-function relationship of TcdB, the secondary structure of the D97 segment was predicted by the algorithms GOR4, SIMPA96, and Chou-Fasman (ProtScal), which showed that the region of AA 1756-1780 of TcdB putatively formed an alpha-helical structure (Figure 1). There is no obvious alpha helical structure formation within the region of AA 1781-1851 according to the predictions. Therefore, we hypothesized that the AA 1756-1780 region of D97 is important. Therefore, 25AA (1756-1780) were deleted to construct the deletion mutant TcdB ∆1756-1780 (Figure 1). The deletion mutant TcdB ∆1827-1851 (Figure 1), with deletion of a different 25AA, 1827-1851, located in the C-terminus of D97, served as the control in the following experiments. The deletion mutants were expressed with His 6 -tags at the C-terminal in a B. megaterium expression system and purified by Ni 2+ affinity chromatography.
To investigate whether the deletion causes improper folding of the proteins, we performed CD spectral analysis and estimated elements of the secondary structure of the toxins on the basis of the CONTIN algorithm. This showed that the mutant toxins maintained a similar secondary structure to TcdB fl (Table 1). The decrease in the alpha helical structure in the composition of TcdB ∆1756-1780 may be due to the deletion of AA 1756-1780.

Cytopathic and Cytotoxic Effects of TcdBΔ1756-1780
The cytopathic effects of TcdBfl, TcdB∆1756-1780, and TcdB∆1827-1851 on CT26 and Vero cell lines were compared using a cell rounding experiment. As shown in Figure 2A, both TcdBfl and TcdB∆1827-1851 caused 100% cell rounding at a concentration of 1 pg/mL, while no obvious cell rounding was observed after treatment with 1 μg/mL TcdB∆1756-1780 for 24 h. Furthermore, the result of the MTT assay indicated that the cytotoxic activity of TcdB∆1756-1780 was decreased by about 6-7 logs compared with TcdBfl and TcdB∆1827-1851 ( Figure 2B). Finally, the in vivo toxicity of the toxins was examined by challenging BALB/C mice, and the survival of the mice was observed ( Figure 2C). Mice challenged with TcdBfl (100 ng/mouse) died within 12 h, while mice challenged with TcdB∆1827-1851 (100 ng/mouse) died within 72 h. By contrast, mice challenged with TcdB∆1756-1780 (100 μg/mouse), a concentration 1000-fold higher than that of TcdBfl and TcdB∆1827-1851, showed no signs of disease and survived until the end of observation (96 h). These results demonstrated that the alpha-helix located at AA 1756-1780 is the critical part of the D97 segment and is essential for the toxicity of TcdB. showing the domain structure of TcdB and the alpha helix, which is located in AA 1756-1780. GTD harbors glucosyltransferase activity. CPD can induce autocleavage of the toxin in the presence of InsP 6 . TD is responsible for GTD domain delivery. RBD is involved in cellular binding and endocytosis. The secondary structure of D97 was predicted using three algorithms (GOR4, SIMPA96, Chou-Fasman (ProtScal)). Alpha-helical and beta-sheet structures are presented as red and green boxes, respectively. The mutant toxins TcdB ∆1756-1780 and TcdB ∆1827-1851 with AA 1756-1780 and 1827-1851 deleted, respectively, are shown in the figure. Abbreviations: GTD, glucosytransferase domain; CPD, cysteine protease domain; TD, translocation domain; HR, hydrophobic region within the translocation domain; RBD, receptor binding domain; D97, the region AA 1827-1851 of TcdB. The cytopathic effects of TcdB fl , TcdB ∆1756-1780 , and TcdB ∆1827-1851 on CT26 and Vero cell lines were compared using a cell rounding experiment. As shown in Figure 2A, both TcdB fl and TcdB ∆1827-1851 caused 100% cell rounding at a concentration of 1 pg/mL, while no obvious cell rounding was observed after treatment with 1 µg/mL TcdB ∆1756-1780 for 24 h. Furthermore, the result of the MTT assay indicated that the cytotoxic activity of TcdB ∆1756-1780 was decreased by about 6-7 logs compared with TcdB fl and TcdB ∆1827-1851 ( Figure 2B). Finally, the in vivo toxicity of the toxins was examined by challenging BALB/C mice, and the survival of the mice was observed ( Figure 2C). Mice challenged with TcdB fl (100 ng/mouse) died within 12 h, while mice challenged with TcdB ∆1827-1851 (100 ng/mouse) died within 72 h. By contrast, mice challenged with TcdB ∆1756-1780 (100 µg/mouse), a concentration 1000-fold higher than that of TcdB fl and TcdB ∆1827-1851 , showed no signs of disease and survived until the end of observation (96 h). These results demonstrated that the alpha-helix located at AA 1756-1780 is the critical part of the D97 segment and is essential for the toxicity of TcdB.

Analysis of Cysteine Protease, Glucosyltransferase Activity, and Cellular Binding of TcdBΔ1756-1780
To determine whether the deletion of AA 1756-1780 influenced the structure and function of CPD, GTD, and RBD, their cysteine protease activity, glucosyltransferase activity, and cellular binding were examined. Autoprocessing by the CPD was tested by an in vitro autocleavage assay. As shown in Figure 3A,B, both TcdB∆1756-1780 and TcdBfl successfully induced autocleavage, releasing a 63kD fragment containing GTD, in the presence of 10 μM InsP6. Furthermore, the concentration dependence of InsP6 and the time course of the cleavage reaction of TcdBΔ1756-1780 were investigated. This showed that TcdBΔ1756-1780 underwent autocleavage after incubation with a series of concentrations of InsP6 for 4, 8, and 12 h. Subsequently, CT26 cell lysate and intact cells were used as substrate, respectively, to check the glucosyltransferase activity of TcdBΔ1756-1780. The TcdBΔ1756-1780 efficiently induced Rac1 glucosylation using CT26 cell lysate as the substrate ( Figure 3D). However, TcdB∆1756-1780 failed to glucosylate Rac1 in intact CT26 cells ( Figure 3C). According to the results above, we concluded that the deletion of amino acids 1756-1780 does not change the structure and function of the cysteine protease and glucosytransferase domains, but speculated that it may result in an inability to deliver GTD into the host cytosol.
Failure of cellular binding, uptake, and translocation may subsequently lead to unsuccessful toxin delivery. The cell surface binding of the mutant toxin was explored using a competition experiment. CT26 cells were incubated with 100 pM TcdBfl in the presence or absence of TcdB∆1756-1780 Groups of mice (n = 6) were injected intraperitoneally with TcdB fl (100 ng/mouse), TcdB ∆1756-1780 (100 µg/mouse), or TcdB ∆1827-1851 (100 ng/mouse). The survival of the mice was monitored for 96 h.

Analysis of Cysteine Protease, Glucosyltransferase Activity, and Cellular Binding of TcdB ∆1756-1780
To determine whether the deletion of AA 1756-1780 influenced the structure and function of CPD, GTD, and RBD, their cysteine protease activity, glucosyltransferase activity, and cellular binding were examined. Autoprocessing by the CPD was tested by an in vitro autocleavage assay. As shown in Figure 3A,B, both TcdB ∆1756-1780 and TcdB fl successfully induced autocleavage, releasing a 63kD fragment containing GTD, in the presence of 10 µM InsP 6 . Furthermore, the concentration dependence of InsP 6 and the time course of the cleavage reaction of TcdB ∆1756-1780 were investigated. This showed that TcdB ∆1756-1780 underwent autocleavage after incubation with a series of concentrations of InsP 6 for 4, 8, and 12 h. Subsequently, CT26 cell lysate and intact cells were used as substrate, respectively, to check the glucosyltransferase activity of TcdB ∆1756-1780 . The TcdB ∆1756-1780 efficiently induced Rac1 glucosylation using CT26 cell lysate as the substrate ( Figure 3D). However, TcdB ∆1756-1780 failed to glucosylate Rac1 in intact CT26 cells ( Figure 3C). According to the results above, we concluded that the deletion of amino acids 1756-1780 does not change the structure and function of the cysteine protease and glucosytransferase domains, but speculated that it may result in an inability to deliver GTD into the host cytosol.
Failure of cellular binding, uptake, and translocation may subsequently lead to unsuccessful toxin delivery. The cell surface binding of the mutant toxin was explored using a competition experiment. CT26 cells were incubated with 100 pM TcdB fl in the presence or absence of TcdB ∆1756-1780 or TcdB CROP (the RBD of TcdB, AA 1852-2366) for 2 h. As observed microscopically ( Figure 3E), a 2500-fold concentration of TcdB ∆1756-1780 competitively inhibited the toxic effect of TcdB fl at a similar level to 5000-fold TcdB CROP . It seemed that the competitive ability of TcdB ∆1756-1780 was two-fold stronger than that of TcdB CROP , which is perhaps due to the existence of a second CROP-independent receptor-binding site and the deletion of region 1756-1780 does not change the binding ability of this CROP-independent binding site. For further confirmation of binding ability of TcdB ∆1756-1780 , CT26 cells were incubated with FITC labeled toxins on ice and imaged by fluorescence microscopy. As shown in Figure 3F, TcdB fl , TcdB ∆1756-1780 , and TcdB ∆1827-1851 were able to bind to the surface of CT26 cells. These results demonstrated that TcdB ∆1756-1780 maintains approximately the same cellular binding activity as full length TcdB. or TcdBCROP (the RBD of TcdB, AA 1852-2366) for 2 h. As observed microscopically ( Figure 3E), a 2500-fold concentration of TcdB∆1756-1780 competitively inhibited the toxic effect of TcdBfl at a similar level to 5000-fold TcdBCROP. It seemed that the competitive ability of TcdB∆1756-1780 was two-fold stronger than that of TcdBCROP, which is perhaps due to the existence of a second CROP-independent receptor-binding site and the deletion of region 1756-1780 does not change the binding ability of this CROP-independent binding site. For further confirmation of binding ability of TcdB∆1756-1780, CT26 cells were incubated with FITC labeled toxins on ice and imaged by fluorescence microscopy. As shown in Figure 3F, TcdBfl, TcdB∆1756-1780, and TcdBΔ1827-1851 were able to bind to the surface of CT26 cells. These results demonstrated that TcdB∆1756-1780 maintains approximately the same cellular binding activity as full length TcdB.

Conformational Change at Low pH and Pore Formation by TcdB∆1756-1780
pH-induced changes in toxin hydrophobicity were identified by the TNS assay, which is a convenient probe for determining the exposure of hydrophobic domains under various conditions. As shown in Figure 4A, both TcdBfl and TcdBΔ1756-1780 exhibited a dramatic increase in TNS-associated fluorescence at pH 4.0. Additionally, when the buffer was neutralized by 1 M NaOH, the fluorescence of both toxins decreased ( Figure 4B-D).  pH-induced changes in toxin hydrophobicity were identified by the TNS assay, which is a convenient probe for determining the exposure of hydrophobic domains under various conditions. As shown in Figure 4A, both TcdB fl and TcdB ∆1756-1780 exhibited a dramatic increase in TNS-associated fluorescence at pH 4.0. Additionally, when the buffer was neutralized by 1 M NaOH, the fluorescence of both toxins decreased ( Figure 4B-D).
used as a negative control. Comparing TcdBΔ1756-1780 with TcdBfl, the fluorescence of TcdBfl increased slightly faster than that of TcdBΔ1756-1780, but this may not be significant because they displayed a similar pattern of increase in general ( Figure 4E). From all these results, we concluded that the mutant toxin TcdBΔ1756-1780 induced conformational change and pore formation under acidic pH conditions in a similar pattern to TcdBfl. However, there is insufficient information about the exact mechanism by which these processes take place. We speculated that the deletion of region 1756-1780 led to locking of the TcdB in endosomes, resulting in a failure to deliver GTD. Total HPTS fluorescence was determined by addition of Triton X-100 (to 0.3%), and His-XynA was used as the negative control. F. Endosome isolation and anti-TcdB western blot. CHO cells were exposed to 2 μg/mL of toxins at 37 °C for 2 h before harvesting. The endosome compartments were isolated by immunoadsorption using anti-Rab5 monoclonal antibody. Subsequently, the isolated endosomes were analyzed by western blot using anti-TcdBCROP antisera.

Behavior in Endosomes
Endosome isolation was performed and western blot analysis was used to obtain further insight into the behavior of the toxins in endosomes. Rab5 is a regulatory guanosine triphosphatase that has been localized to the plasma membrane, clathrin-coated vesicles, and early endosomes. It participates Total HPTS fluorescence was determined by addition of Triton X-100 (to 0.3%), and His-XynA was used as the negative control. F. Endosome isolation and anti-TcdB western blot. CHO cells were exposed to 2 µg/mL of toxins at 37 • C for 2 h before harvesting. The endosome compartments were isolated by immunoadsorption using anti-Rab5 monoclonal antibody. Subsequently, the isolated endosomes were analyzed by western blot using anti-TcdB CROP antisera.
Exposure of the hydrophobic region of the toxins enabled membrane insertion, which is paralleled by pore formation, resulting in the entrance of toxins into host cells directly through the endosomal membrane. The pore formation of TcdB ∆1756-1780 was studied by assay of fluorophore leakage from HPTS/DPX loaded LUVs. Pore formation was monitored by the release of the fluorescent dye HPTS, which is quenched in the lipid vesicles by DPX. The HPTS/DPX compounds were released and diluted when pores formed in the LUVs, which caused an increase in fluorescence by reduction of collision quenching. As shown in Figure 4E, under low pH conditions, TcdB fl and TcdB ∆1756-1780 induced pore formation in LUVs and caused an increase in fluorescence. However, no increase in the fluorescent signal was observed under acidic pH with protein His 6 -XynA, which was used as a negative control. Comparing TcdB ∆1756-1780 with TcdB fl , the fluorescence of TcdB fl increased slightly faster than that of TcdB ∆1756-1780 , but this may not be significant because they displayed a similar pattern of increase in general ( Figure 4E). From all these results, we concluded that the mutant toxin TcdB ∆1756-1780 induced conformational change and pore formation under acidic pH conditions in a similar pattern to TcdB fl . However, there is insufficient information about the exact mechanism by which these processes take place. We speculated that the deletion of region 1756-1780 led to locking of the TcdB in endosomes, resulting in a failure to deliver GTD.

Behavior in Endosomes
Endosome isolation was performed and western blot analysis was used to obtain further insight into the behavior of the toxins in endosomes. Rab5 is a regulatory guanosine triphosphatase that has been localized to the plasma membrane, clathrin-coated vesicles, and early endosomes. It participates in endosomal membrane fusion reactions and is important in control of endocytic function [24,25]. Rab5 is often used as an early endosome marker [24]. In this study, endosome populations were purified by immunoadsorption from syringe-lysed cell lysate using an antibody against Rab5 protein and were analyzed by western blot using anti-TcdB antibodies. The result showed ( Figure 4F) that holotoxin of TcdB fl and TcdB ∆1827-1851 can be recognized by anti-TcdB CROP antibody, which is specific for the receptor binding domain of TcdB, but the band of TcdB ∆1827-1851 was much weaker than that of TcdB fl . In contrast, a very small amount the holotoxin of TcdB ∆1756-1780 was detected as well as several bands of lower molecular weight. Based on these results, we hypothesize that the majority of TcdB ∆1756-1780 is rapidly degraded in the endosome, which prevents efficient delivery of GTD into the cytosol. Therefore, a much higher concentration of TcdB ∆1756-1780 would be needed to induce the same level of cell rounding or cell death when compared with wild type TcdB.

Discussion
The goal of this study is to identify the critical part of D97 segment and provide insight into the structure-function relationship of the TcdB translocation action. Over decades, efforts have been made to reveal the mechanism of delivery of the cytotoxic glucosyltransferase domain across the endosomal membrane into cytosol. Despite significant advances, there is still a long way to completely reveal the underlying mechanism of the toxin delivery. In 2013, we found that a 97-amino-acid segment (D97) located in the C-terminus of the translocation domain is essential for the toxicity of TcdB, and the D97 segment was hypothesized to be involved in the translocation of TcdB because TcdB-D97 failed to release GTD into the host cytosol [23]. In this study, we narrowed down the critical region of D97 and tried to investigate the function of this region. Using algorithms GOR4, SIMPA96, and Chou-Fasman (ProtScal), the secondary structure of D97 segment was predicted, demonstrating that the region covering AA 1756-1780 formed an alpha helical structure and it was hypothesized to be important. Therefore, a TcdB deletion mutant TcdB ∆1756-1780 was constructed, while TcdB ∆1827-1851 was used as a control protein. TcdB ∆1827-1851 still exhibited almost complete in vitro and in vivo toxicity when compared with TcdB fl , however, toxicity of TcdB ∆1756-1780 had dramatically decreased by 6-7 logs. To this point, it is inferred that the region 1756-1780 plays an important role in D97, while the region 1781-1851 may not be essential for the toxicity of TcdB. To investigate the function of the region 1756-1780, the glucosyltransferase and cysteine protease activities, cellular binding, pH-dependent conformational change and pore formation of TcdB ∆1756-1780 were examined. As results, TcdB ∆1756-1780 maintains approximately intact functions of the GTD, CPD, and RBD as TcdB fl . However, TcdB ∆1756-1780 was able to induce Rac1 glucosylation only when cell lysate was used as the substrate but failed when using intact cells as substrate. It is hypothesized that the deletion of 25AA 1756-1780 impairs the function of the TD. Previous studies showed that the translocation process of C. difficile TcdB is associated with pH-induced conformational change and pore formation. Therefore, pH-induced conformational change and pore formation were studied by TNS fluorescence analysis and the LUVs fluorophore leakage assay respectively. The results indicated that TcdB ∆1756-1780 had a similar pattern of conformation change and pore formation to TcdB fl under acidic conditions, which would suggest similar function in endosomes. According to our previous report, TcdB-D97 was concluded to be trapped in endosomes. Therefore, we investigated the differences between TcdB fl and TcdB ∆1756-1780 related to the behavior in endosomes. We found that TcdB ∆1756-1780 was degraded rapidly in endosomes into several fragments with different molecular weights. It is important to note that the region 1756-1780 is only a small part within the translocation domain, the deletion of which may not influence the general trend of conformational change, that is why TcdB ∆1756-1780 displayed conformational change in similar pattern as TcdB fl in the TNS fluorescence analysis. Nevertheless, the region 1756-1780 may be essential for a critical step in conformational change, and its deletion led to unsuccessful membrane insertion and eventually prevented GTD from being translocated across the endosomal membrane.
Recently, Björn Schorch et al. [14] proposed a two-receptor model for the cell entry of clostridial glycosylating toxins, in which the CROP domain primarily facilitates the accumulation of the toxins at the cell surface, and the additional receptor-binding domain interacts with another specific cell surface protein to induce the endocytosis of the toxins. A new model of the modular composition of TcdB presented by Selda Genisyuerek et al. [16] suggests that the region between AA 1500 and 1851 is not necessary for translocation and may represent an additional receptor-binding site, as deletion of the region reduced toxicity, but did not complete abrogate it. TcdB ∆1756-1780 was detected in the endosome, which indicates that it is able to induce endocytosis. Therefore, we suggest that the region 1756-1780 is not necessary for cellular binding but is necessary for efficient GTD translocation. Because of ineffective translocation, after endocytosis, TcdB ∆1756-1780 was trapped in the endosome and was subsequently degraded into small fragments. The majority of the intact GTD was therefore unable to be translocated into the host cytosol. As a result, a much higher concentration of TcdB ∆1756-1780 would be needed to trigger same level of cell rounding or cell death as for TcdB fl .
In summary, we identified a region, 25AA 1756-1780, within D97 that is essential for TcdB toxicity. The region 1756-1780 plays a pivotal role in the translocation of GTD across the endosomal membrane, however, it is not necessary for pore formation and cellular binding. We hypothesize that the region 1756-1780 has a critical role in pH-induced conformational change, and its deletion may lead to an incorrect conformational change, triggering steric hindrance, consequently resulting in unsuccessful membrane insertion and GTD delivery. Eventually, deletion of region 1756-1780 leads to toxin trapped and degraded in the endosome compartment ( Figure 5). Determination of the crystal structure and further study of the translocation domain are needed to explain the function of the 1756-1780 region and to clarify the mechanism of action involved in TcdB translocation. Genisyuerek et al. [16] suggests that the region between AA 1500 and 1851 is not necessary for translocation and may represent an additional receptor-binding site, as deletion of the region reduced toxicity, but did not complete abrogate it. TcdB∆1756-1780 was detected in the endosome, which indicates that it is able to induce endocytosis. Therefore, we suggest that the region 1756-1780 is not necessary for cellular binding but is necessary for efficient GTD translocation. Because of ineffective translocation, after endocytosis, TcdB∆1756-1780 was trapped in the endosome and was subsequently degraded into small fragments. The majority of the intact GTD was therefore unable to be translocated into the host cytosol. As a result, a much higher concentration of TcdB∆1756-1780 would be needed to trigger same level of cell rounding or cell death as for TcdBfl. In summary, we identified a region, 25AA 1756-1780, within D97 that is essential for TcdB toxicity. The region 1756-1780 plays a pivotal role in the translocation of GTD across the endosomal membrane, however, it is not necessary for pore formation and cellular binding. We hypothesize that the region 1756-1780 has a critical role in pH-induced conformational change, and its deletion may lead to an incorrect conformational change, triggering steric hindrance, consequently resulting in unsuccessful membrane insertion and GTD delivery. Eventually, deletion of region 1756-1780 leads to toxin trapped and degraded in the endosome compartment ( Figure 5). Determination of the crystal structure and further study of the translocation domain are needed to explain the function of the 1756-1780 region and to clarify the mechanism of action involved in TcdB translocation.

CT26 cells (BALB/C mouse colon tumor cells), Vero cells (kidney epithelial cells from
African green monkeys), and CHO cells (Chinese hamster ovary cells), purchased from the Chinese Academy of Sciences Institute of Cell Resource Center, were cultured in Dulbecco's minimum Eagle's medium supplemented with 10% fetal bovine serum, 100 units/mL penicillin, and 100 µg/mL streptomycin in 100 mm culture plates at 37 • C and 5% CO 2 .

Protein Expression and Purification
The transformation of B. megaterium protoplast was performed according to the manufacturer's instructions (MoBiTec, Goettingen, Germany). The transformed B. megaterium colonies were transferred to the LB broth medium containing 10 µg/mL tetracycline and incubated overnight at 37 • C with 250 rpm. The overnight culture was diluted 1:100 in LB broth medium containing tetracycline and grown to an optical density OD 600 around 0.3 before the addition of xylose (5 mg/mL) to induce protein expression.
All proteins were expressed with C-terminal His 6 tags. The purification of His-tag proteins was performed by Ni 2+ affinity chromatography. Briefly, the B. megaterium pellet was suspended in 5 mL lysis buffer (20 mM phosphate sodium buffer, 500 mM NaCl, 30 mM imidazole, pH 7.4) per 100 mL bacterial culture. Cells were disrupted by sonication and the lysate was centrifuged at 15,500× g for 30 min at 4 • C. The supernatant was applied to a nickel-charged HisTrap HP column (GE Healthcare Bio-Sciences, Pittsburgh, PA, USA) and the bound protein was eluted using elution buffer (20 mM phosphate buffer, 500 mM NaCl, 500 mM imidazole, pH 7.4). The proteins were dialyzed to PBS buffer containing 20% glycerol and stored at −80 • C.
Circular dichroism (CD) spectrophotometry was performed using a Chirascan Circular Dichroism Spectropolarimeter (Applied Photophysis Limited, Surrey, UK), with a scan interval of 1 nm and path length of 0.1 cm at 25 • C. The CD data were analyzed on the basis of the CONTIN algorithm.

Cytotoxic and Cytopathic Effects
The cytotoxic effect induced by the toxins was analyzed by MTT (Methylthiazolyldiphenyltetrazolium bromide) viability assay as described previously [26]. Briefly, 2 × 10 4 CT26 cells were seeded in a 96-well plate and cultured at 37 • C for 24 h. Serial dilutions of each toxin were added to the cells and incubated at 37 • C for another 72 h. Following this, 10 µL of MTT was added and the plate was further incubated at 37 • C for 2 h. The formazan was solubilized with DMSO, and absorbance at 570 nm was measured using a SpectraMax M5 (Molecular Devices, Sunnyvale, CA, USA). Cell viability was expressed as a percentage of the cell survival in the control wells.
The cytopathic effect was measured by a cell rounding experiment. CT26 or Vero cells were incubated with TcdB fl (full length TcdB), TcdB ∆1756-1780 , and TcdB ∆1827-1851 at indicated concentrations for 24 h and cell rounding was observed by light microscopy. The experiments were repeated three times, and triplicate wells were assessed for the MTT assay and cell rounding in each experiment.

Mouse Systemic Toxin Challenge
6-to 8-week-old BALB/C mice (SPF) were purchased from Guangdong Medical Laboratory Animal Center (Guangdong, China). Groups of mice (n = 6) were challenged intraperitoneally with TcdB fl (100 ng/mouse), TcdB ∆1756-1780 (100 µg/mouse), and TcdB ∆1827-1851 (100 ng/mouse), respectively. Mouse survival was monitored every six hours. The animal protocols used in this work were approved by Guangdong Provincial Department of Science and Technology (Approval Number: SYXK (Yue) 2014-0145). This research does not violate any national guidelines and institutional policies for use of animal in research.

In Vitro Autocleavage Assay
The in vitro autocleavage assay was performed as described previously [6,27]. Each toxin protein was diluted in 20 mM Tris buffer (pH 7.4) in a final volume of 100 µL. Cleavage was initiated by addition of 10 µM inositol hexakisphosphate (InsP 6 ) (Sigma, St. Louis, MO, USA) and the mixture was incubated at 37 • C for 12 h. To investigate the dependence of the effect on the time course and the concentration of InsP 6 , the toxins were incubated with 5, 10, and 20 µM InsP 6 for 4, 8, and 12 h, respectively. The reaction was stopped by SDS-PAGE sample loading buffer, and analyzed by western blot using anti-TcdB GTD (the GTD of TcdB, amino acids 1-543) antiserum prepared by our laboratory.

In Vitro Glucosylation Assay
The in vitro glucosylation assay was performed with intact CT26 cells and cell lysates. In the experiment with intact cells, 2 × 10 5 CT26 cells were seeded in a 24-well plate and cultured at 37 • C in 5% CO 2 for 36 h before being treated with different concentrations of TcdB fl , TcdB ∆1756-1780 , and TcdB ∆1827-1851 respectively for 4 h. After treatment, the cells were washed with PBS three times, then lysed by SDS-PAGE sample loading buffer and boiled for 5 min. In the cell lysate experiment, CT26 cell pellets were resuspended in a reaction buffer (50 mM HEPES pH 7.5, 100 mM KCl, 1 mM MnCl 2 , and 2 mM MgCl 2 ) and lysed by passing through a 30 G needle 40 times. After centrifugation, the supernatant was used as the cytosolic fraction. For the glucosylation assay, the cytosolic fraction was incubated with 100 µg/mL TcdB fl or TcdB ∆1756-1780 at 37 • C for 4 h. The reaction was terminated by adding SDS-PAGE sample loading buffer and was boiled for 5 min. Finally, the samples were analyzed by anti-unglucosylated Rac1 mAb (BD Biosciences, San Diego, CA, USA) using western blot analysis.

Analysis of Cell Surface Binding of the Toxins
In the competition experiment, CT26 cells were incubated with 10 pM TcdB fl for 3 h in the presence or absence of 250 µM TcdB ∆1756-1780 , or 500 µM TcdB CROP (the RBD of TcdB, amino acids 1852-2366). Meanwhile, CT26 cells were incubated with 0.5 nM TcdB 1-1851 (the fragment without CROP, amino acids 1-1851) for 4 h with or without 250 µM TcdB ∆1756-1780 . Subsequently, the morphology of the CT26 cells was visualized by light microscopy.