Cell penetration of bacterial protein toxins

Abstract

I mportant advances have been made recently in the study of cell penetration by bacterial protein toxins . I,* Many protein toxins from bacteria and plants have their effects in the cytosol of cells. Rather than reaching their targets directly through the plasma membrane, these toxins exploit the endocytotic pathway and enter the cytosol from intracellular compartments. As discussed recently3, the process of cell intoxication consists of four steps: (1) binding, (2) internalization, (3) membrane translocation, and (4) target modification. This elaborate mechanism of action relies on a common structural architecture. These toxins are all made up of two main elements linked via a single disulfide bridge, A being an enzymatic subunit and B a protomer that mediates cell binding and penetration. The B portions of some toxins, including cholera toxin, Escherichia coli heatlabile toxin, Shiga toxins and pertussis toxin, are oligomers that bind lipidor protein-linked sugars of the cell surface (Refs 4,5 and references therein). Another group of toxins consists of diphtheria toxin (DT), exotoxin A of Pseudomonas aeruginosa (ETA), tetanus (TeNT) and botulism neurotoxins and the Bacillus anthracis toxic complex [protective antigen (PA), edema factor and lethal factor]. Their B portions are made up of two domains: a receptor-binding part, linked to another domain involved in membrane translocation”~“,6-8. These latter toxins are thus threedomain toxins. After cell binding, protein toxins are internalized inside vesicles and, according to their specificity, enter different pathways of intracellular membrane traffic. A large body of evidence indicates that the threedomain toxins end up inside acidic intracellular compartments, and that the low pH is instrumental in causing the entry of the catalytic domain A into the cytosol (see Ref. 3 for references). The reports of Eriksen et al.’ and of Beise et al.’ deal with membrane translocation, the least understood of the four steps of the whole intoxication process. It has long been known that DT forms ion channels across planar lipid bilayers at low pH (Refs 9,lO) and that this property is shared by all three-domain toxins1’-‘4. These channels are formed by the B portion, the amino-terminal domain of which plays the major role. This was the first evidence that these toxins change conformation at low pH in such a way that the B part inserts into the lipid bilayer. Moreover, these studies introduced a very sensitive experimental system for measuring membrane penetration and translocation of toxins. However, such an approach has two potential biases: (1) an ‘artificial’ membrane is used, and it is not clear to what extent results can be extrapolated to the in viva situation; (2) the conductance of toxin ion channels in planar lipid bilayers (a few pS) is much lower than that of a bona fide protein-translocating pore (hundreds of pS)“. More recently, DT and PA have been found to form ion channels in living cells’x-lx; however, their electrical properties have not been studied. Now, Eriksen et al.’ and Beise et al.’ have applied the patch-clamp technique to cells exposed to DT and TeNT, and have been able to demonstrate that these toxins do form discrete channels with con-