Clostridium difficile infections can induce mild to severe diarrhoea and the often associated characteristic pseudomembranous colitis. Two protein toxins, the large glucosyltransferases TcdA and TcdB, are the main pathogenicity factors that can induce all clinical symptoms in animal models. The classical molecular mode of action of these homologous toxins is the inhibition of Rho GTPases by mono-glucosylation. Rho-inhibition leads to breakdown of the actin cytoskeleton, induces stress-activated and pro-inflammatory signaling and eventually results in apoptosis of the affected cells. An increasing number of reports, however, have documented further qualities of TcdA and TcdB, including the production of reactive oxygen species (ROS) by target cells. This review summarizes observations dealing with the production of ROS induced by TcdA and TcdB, dissects pathways that contribute to this phenomenon and speculates about ROS in mediating pathogenesis. In conclusion, ROS have to be considered as a discrete, glucosyltransferase-independent quality of at least TcdB, triggered by different mechanisms.

Clostridioides difficile toxins TcdA and TcdB are large clostridial glucosyltransferases which are the main pathogenicity factors in C. difficile-associated diseases. Four highly conserved cysteines are present in all large clostridial glucosyltransferases. In this study we focused on the conserved cysteine 2232 within the combined repetitive oligopeptide domain of TcdB from reference strain VPI10463 (clade I). Cysteine 2232 is not present in TcdB from hypervirulent strain R20291 (clade II), where a tyrosine is found instead. Replacement of cysteine 2232 by tyrosine in TcdBV PI10463 reduced binding to the soluble fragments of the two known TcdB receptors, frizzled-2 (FZD2) and poliovirus receptor-like protein-3/nectin-3 (PVRL3). In line with this, TcdBR20291 showed weak binding to PVRL3 in pull-down assays which was increased when tyrosine 2232 was exchanged for cysteine. Surprisingly, we did not observe binding of TcdBR20291 to FZD2, indicating that this receptor is less important for this toxinotype. Competition assay with the receptor binding fragments (aa 1101–1836) of TcdBV PI10463 and TcdBR20291, as well as antibodies newly developed by antibody phage display, revealed different characteristics of the yet poorly described delivery domain of TcdB harboring the second receptor binding region. In summary, we found that conserved Cys-2232 in TcdB indirectly contributes to toxin–receptor interaction.

Toxin A and Toxin B (TcdA/TcdB) are large glucosyltransferases produced by Clostridium difficile. TcdB but not TcdA induces reactive oxygen species-mediated early cell death (ECD) when applied at high concentrations. We found that nonglucosylated Rac1 is essential for induction of ECD since inhibition of Rac1 impedes this effect. ECD only occurs when TcdB is rapidly endocytosed. This was shown by generation of chimeras using the trunk of TcdB from a hypervirulent strain. TcdB from hypervirulent strain has been described to translocate from endosomes at higher pH values and thus, meaning faster than reference type TcdB. Accordingly, intracellular delivery of the glucosyltransferase domain of reference TcdB by the trunk of TcdB from hypervirulent strain increased ECD. Furthermore, proton transporters such as sodium/proton exchanger (NHE) or the ClC-5 anion/proton exchanger, both of which contribute to endosomal acidification, also affected cytotoxic potency of TcdB: Specific inhibition of NHE reduced cytotoxicity, whereas transfection of cells with the endosomal anion/proton exchanger ClC-5 increased cytotoxicity of TcdB. Our data suggest that both the uptake rate of TcdB into the cytosol and the status of nonglucosylated Rac1 are key determinants that are decisive for whether ECD or delayed apoptosis is triggered.

Binary toxins are produced by several pathogenic bacteria. Examples are the C2 toxin from Clostridium botulinum, the iota toxin from Clostridium perfringens, and the CDT from Clostridium difficile. All these binary toxins have ADP-ribosyltransferases (ADPRT) as their enzymatically active component that modify monomeric actin in their target cells. The binary C2 toxin was intensively described as a tool for intracellular delivery of allogenic ADPRTs. Here, we firstly describe the binary toxin CDT from C. difficile as an effective tool for heterologous intracellular delivery. Even 60 kDa glucosyltransferase domains of large clostridial glucosyltransferases can be delivered into cells. The glucosyltransferase domains of five tested large clostridial glucosyltransferases were successfully introduced into cells as chimeric fusions to the CDTa adapter domain (CDTaN). Cell uptake was demonstrated by the analysis of cell morphology, cytoskeleton staining, and intracellular substrate glucosylation. The fusion toxins were functional only when the adapter domain of CDTa was N-terminally located, according to its native orientation. Thus, like other binary toxins, the CDTaN/b system can be used for standardized delivery systems not only for bacterial ADPRTs but also for a variety of bacterial glucosyltransferase domains.