During the past decade an increasing incidence of C. difficile disease has been evident worldwide. This bacterium is now recognized as a major nosocomial pathogen in industrialized parts of the world (Lyerly and Wilkins, 1995). The pseudomembranous colitis caused by C. difficile was early recognized as a toxin disease, i.e., all symptoms can be evoked by the toxins. Understanding the disease thus requires an understanding of the toxins.
Our knowledge of the molecular toxicology of C. difficile has been considerably improved as the enzymatic mode of action of the toxins have been identified. This knowledge also has the potential to help in answering a whole array of questions related to the functions of small G-proteins in cells. However, a large number of obvious questions remain both concerning the toxins perse, and their actions on single cells and in the intestine.
One intriguing point is the significant differences in cytotoxic and enterotoxic potencies between toxins A and B. What determines these comparison differences? Could the difference in cytotoxic potency simply reflect a of Tox A and Tox B difference in the affinity of the toxins for their cellular substrates? Preliminary observations in our laboratory suggest that ToxA, in gluco-sylation assays in vitro, using either cell lysates or recombinant Rho, is much less potent than ToxB with regard to glucosylation potency (Chaves Olarte et a/., unpublished). However, the toxins obviously also have different cell surface receptors, and this could be part of the explanation. Despite the high homology at the primary sequence level, the native 3D-structures certainly differ. This is evident from the different epitope patterns, and also from the finding thatToxA is highly resistant to trypsin whereas ToxB is inactivated by this enzyme (Lyerly et a/., 1989). The uneven epitope distribution on ToxA may imply that the native toxin is more or less covered by its C-terminal repeat structure, possibly by being dimerized in a manner that exposes predominantly the repeat structure. In contrast, ToxB is likely to have a 3D-configuration that also exposes some N-terminal parts of the molecule, as its epitopes appear to be distributed over the entire primary sequence of the molecule. Establishment of the 3D-structures for both toxins will be helpful in solving the receptor problem, and will answer several other questions as well.
In the cellular intoxication steps occurring after receptor binding, both toxins require a low pH compartment and some kind of processing to be optimally active in the cytosol. The tertiary structures are likely to be unfolded at low pH, possibly exposing the common hydrophobic domain, which might promote toxin translocation across the endosomal membrane and delivery of (at least) the catalytically active part of the molecules to the cytosol. Nothing is known about how the toxins are processed in cells, or concerning the mode of toxin translocation across membranes. The 3D-structures created after these events most probably differ from the native ones. They should allow mapping of the amino acids forming the catalytic site, including the region(s) required for binding to the target GTPases as well as for binding the co-substrate UDP-Glc. Preparing fragments of the toxins for microinjection into cells will help to solve these latter questions, and also indicate whether the lower glucosylation potency of ToxA is due to a real difference in enzymatic potency or to a partial masking of the catalytic site by the C-terminal repeat structure.
Another question to be explored is whether the C-terminal repeat regions alone, given extracellularly, could evoke some of the biological effects that have been described. For instance, an 873 amino acid repeat from ToxA can apparently elicit a rapid increase in cyto-solic calcium in Chinese hamster ovary cells overexpressing sucrase isomaltase, a glycoprotein suggested as an intestinal receptor for ToxA (Thomas LaMont, personal communication). Would this same part of the ToxA molecule elicit fluid secretion in the intestine, explaining the postulated signaling capacity of ToxA? Does the ToxB repeat in isolation have any biological activity? Furthermore, high amounts of ToxB consistently cause a rapid "lytic" effect on cells, which has been correlated with a phospholipase A2 activation (Shoshan et a/., 1993a). Since this phenomenon also occurred in the toxin-resistant UDP-Glc-deficient mutant cell, it is unrelated to the cytoskeletal effect mediated by Rho-glucosylation. How is the phospholipase A2 activation generated and does it have any implications for disease induced by C. difficile? Some of the answers to these questions may help to clarify the true events in intestinal pathogenesis, and perhaps also give a clue to the intriguing phenomenon that neonates are resistant to C. difficile toxins in the intestine. Equally intriguing is the question of why and how animals die upon parenteral injection of low amounts of the toxins (Arnon et al., 1984; Lyerly et al., 1986), and particularly the fact that the differences in enterotoxic and cytotoxic potencies are apparently not reflected in the lethal activities of the toxins. Defined fragments of the toxins will also be useful for clarifying this aspect of their biological actions.
Obviously the C. difficile toxins will become extremely useful tools applications for further studies of how the Rho subfamily proteins control the of C. difficile toxins ACTSK, as well as aspects of cell proliferation in different types of cells (Olson et al., 1995). These toxins will be easier to use for manipulation of small GTPases than the exoenzyme C3 from C. botulinum, because they are internalized into cells and are more potent. Their drawback compared with C3 is that they attack more than one target protein. Preparation of mutant toxins, fragments of toxins or hybrid toxins which discriminate between various GTPase targets might solve this problem and allow sophisticated studies of the cellular cross-talk between small GTPases (see further discussion in Chapters 10 and 15).
The use of the C. difficile toxins as tools should obviously not be restricted to basic cell biological applications. The possibility of using the toxins, or fragments of them, for vaccination purposes should be investigated. Moreover, it is tempting to suggest that potent cytotoxins like these might be useful in targeted tumor therapy. The recent study of Riegler and coworkers (1995) indicated that 3 nM ToxA had no effect on human colonic mucosal sheets, at least during a 5h exposure. On the other hand as little as 0.01 -0.001 nM ToxA was shown to kill a variety of human colonic and pancreatic tumor cell lines, while 12-500 times more toxin was needed to kill cell lines derived from normal tissues (Kushnaryov etal., 1992). Is there a possibility of using ToxA locally as an antitumor agent directed against colon cancer?
Was this article helpful?