How a Bacterial Toxin Causes Diarrhea
By: Carol A. Rouzer, VICB Communications
Published: August 6, 2010
VICB investigators show how the structure of the Clostridium difficile major toxins enables them to enter cells.
The bacterium Clostridium difficile (Figure 1) has become the major cause of hospital-acquired diarrhea, including a particularly virulent form of infection known as pseudomembranous colitis. An anaerobic gram-positive bacillus (rod-shaped bacterium), C. difficile can frequently be found among the intestinal flora of healthy individuals, but its growth is normally suppressed by the presence of overwhelming numbers of non-pathogenic bacteria. This balance is upset by the use of broad-spectrum antibiotics, which kill the harmless bacteria allowing an overgrowth of C. difficile. Elderly and debilitated patients are particularly at risk.
Figure 1. Scanning electron micrograph of C. difficile obtained from a human stool specimen. (Image obtained courtesy of Wikimedia Commons under the GNU Free Documentation License.
Like many pathogenic bacteria, C. difficile causes illness by secreting toxins which damage cells and invoke an inflammatory response. The two major toxins of C. difficile, TcdA and TcdB, are large proteins that damage host cells via a multistep process, each step of which involves a separate portion of the molecule. First, a binding domain allows the toxin to attach to sugar-containing proteins on the cell’s surface. Once this occurs, the toxin enters the cell through the process of endocytosis which directs it to the acidic endosomal compartment of the cell. The acid in the endosomes causes the toxin molecule to change shape, forming an elongated structure that enables the toxin to form a pore through the endosomal membrane. Next a protease domain cleaves the toxin’s amino acid chain, freeing a portion of the protein called the glucosyltransferase domain. This portion of the toxin leaves the endosome through the pore, thereby reaching the cytosol where it transfers sugars to a class of cell signaling proteins known as Rho GTPases. The sugar addition inactivates these proteins, resulting in abnormal intra- and extra-cellular communications, ultimately leading to inflammation or cell death (Figure 2).
Figure 2. Healthy CHO cells (left) have an elongated shape. After 4 h exposure to 2 μg of TcdA (center) or TcdB (right) the cells round up and detach from the culture dish, indicating cell death. Used with permission from Pruitt et al. (2010) Proc. Natl. Acad. Sci. U.S.A., published online July 12, DOI: 10.1073/pnas.1002199107. Copyright 2010.
The proposed mechanism of action for the C. difficile toxins suggests a complex structure consisting of separate molecular regions each designed for a distinct function. Although past studies have characterized the three-dimensional structure of separate portions of the molecule, an understanding of how these regions are integrated to form the complete toxin protein has remained a mystery until VICB member Borden Lacy, and collaborator Melanie Ohi applied the technique of negative stain electron microscopy to obtain pictures of individual TcdA and TcdB molecules [Pruitt et al. (2010) Proc. Natl. Acad. Sci. U.S.A., published online July 12, DOI: 10.1073/pnas.1002199107].
In negative stain electron microscopy, protein molecules are dispersed onto a carbon-coated grid, so that distinct, individual molecules can be visualized. Coating the proteins with a high molecular weight stain, such as the uranyl acetate stain used by the Lacy and Ohi labs coats, provides a contrast medium that deflects the microscope’s electron beam differently from the protein. The result is a high contrast image of individual protein molecules. Analysis of the shapes of a large number of molecules allows mathematical averaging to obtain a representative image of the protein’s shape (Figure 3).
Figure 3. Example of negative stain electron micrographs of TcdA (top) and TcdB (bottom). In each case the large figure shows the appearance of the raw data with individual proteins circled. The small inset shows the results of mathematical averaging that reveals the shape of the protein. Used with permission from Pruitt et al. (2010) Proc. Natl. Acad. Sci. U.S.A., published online July 12, DOI: 10.1073/pnas.1002199107. Copyright 2010.
Initial studies indicated that the structures of TcdA and TcdB were similar, so the investigators concentrated on defining the structure of TcdA, which showed greater uniformity. They found that the protein is comprised of a head consisting of two globular “pincher-like” domains, a long kinked tail extending from the bottom of the larger of the two pinchers, and a shorter tail connected to the smaller of the pinchers. The investigators were able to show that the long tail corresponds to the sugar binding domain, the head is the pore-forming domain, and the short tail is the glucosyltransferase domain (Figure 4A). The proposed model correlated well with previous X-ray crystallographic data, as indicated by the fact that structures based on those data could be neatly docked by computer modeling in to the outlines of the model (Figure 5).
Figure 4. Model of TcdA under neutral (A) and acidic (B) conditions. The long tail (green) bind the toxin to the target cell. The head domain (yellow) forms the membrane pore. The protease domain (blue) cleaves to protein to release the short tail (red) which transfers sugar groups to Rho GTPase target proteins. Used with permission from Pruitt et al. (2010) Proc. Natl. Acad. Sci. U.S.A., published online July 12, DOI: 10.1073/pnas.1002199107. Copyright 2010.
Figure 5. Overlay of crystal structure data with proposed model for the binding domain (green), protease domain (blue), and glucosyltranferase domain (red). Used with permission from Pruitt et al. (2010) Proc. Natl. Acad. Sci. U.S.A., published online July 12, DOI: 10.1073/pnas.1002199107. Copyright 2010.
Having established the shape and configuration of TcdA, the Lacy and Ohi labs repeated their experiments under acidic conditions to reflect exposure of the toxin to the environment of the endosome. They found that the head domain of the protein exhibited a marked change in shape, exemplified by elongation of the large pincher region (Figure 4B). This shape change correlates very well with the pore-forming function of the head domain of the protein. Thus, the model proposed by the Lacy and Ohi labs provides a molecular mechanism for TcdA intoxication. TcdA and TcdB belong to the family of large clostridial toxins which share many of the same structural features. Therefore, it is highly likely that this model provides critical information for understanding the mechanism of all of these important disease-causing proteins.