Protective Antigen: Moving Edema Factor and Lethal Factor Into Cells

Although PA is a component of anthrax toxin, by itself it is not toxic.  The earliest indication of its function came from a series of experiments performed by Molnar and Altenbern [48] using partially purified toxin components.  They noted that PA was removed from the circulation within 1-2 h of its injection into Fischer rats.  By contrast, LF remained in the bloodstream for at least 4 h following its injection.  If LF or PA+LF was injected shortly after PA, then the injected animal died within 70 min.  However, if LF or PA+LF were injected 1 h after PA, then the time to death after injection of the toxic component was delayed.  Moreover, if LF was injected more than 2 h after PA it exhibited no toxicity.  A similar result was obtained if PA+LF was injected 4 h after PA.  These results indicate that PA is bound and sequestered by a component of the host’s tissues and that, in order to exert its toxic effects, LF must interact with PA prior to its sequestration.  This mode of action closely resembles that of A-B binary toxins in which the component having the enzymatic activity (“A” domain) gains entry to the cytosol via a cell binding component (“B”) [49].  Thus, PA is envisioned to bind a cell surface receptor and translocate LF (or EF; see below) to the cytosol (Fig. 1).  Direct evidence for this hypothesis was not obtained until 1982 when Leppla [34] showed that PA was required for entry of EF into the cytosol of Chinese hamster ovary cells.

Before it is able to translocate EF or LF to the cytosol, PA must first be proteolytically modified (Fig. 1).  Cleavage occurs at the sequence RKKR¹⁶⁷, resulting in the removal of a 20-kDa fragment, likely by furin or a similar cell surface–associated protease [14,50], yielding PA63.  Deletions of or mutations at this site render PA resistant to proteolysis and consequently nontoxic in combination with LF or EF [50,51].  Removal of the NH₂-terminal fragment is apparently necessary to expose a region of PA that can bind to the other toxin components, since uncleaved PA can bind the cell but is unable to associate with EF or LF [50,51].  Analysis of COOH terminal truncations of PA63 indicate that the C-terminus is necessary for cell binding [26].  Similarly, using fragments of PA generated by chymotrypsin and trypsin treatment, Novak et al. [52] were also able to show that the C-terminus of PA was necessary for cell binding.

PA63 can form cation-selective channels in planar phospholipid bilayers whereas intact PA cannot [53].  Observations by nondenaturing gel electrophoresis [53] and electron microscopy [54] indicate that seven molecules of PA63 are required to form each channel [54].  More recently, X-ray crystallographic analysis of PA shows that heptamerization is sterically blocked until the removal of the NH₂-terminal 20-kDa fragment [13].

Several experiments indicate that PA mediates entry of LF and EF into the cell via the endosomal pathway.  Friedlander [9] showed that by pre-exposing mouse peritoneal macrophages to agents that can dissipate intracellular proton gradients (such as ammonium chloride), it was possible to render the macrophages resistant to lethal toxin.  This effect was not caused by a block to toxin internalization since bound PA was not accessible to antibodies.  Subsequent exposure of these cells to an acidic environment could reverse this inhibition.  Also, the uptake of EF by cells can be blocked by treatment with the microfilament inhibitor cytochalasin D, and this effect may also be overcome via acidification [55].  Thus, it may be concluded that lethal toxin must pass through an acidic, endocytic vesicle to exert its toxic effects within the cytosol.

Figure 1.  An outline of LeTx activity.  Protective antigen (PA) binds an anthrax toxin receptor (ATR) on the cell surface where it is proteolytically modified by a furin-like protease leaving PA₆₃ still bound to the cell.  This allows PA₆₃ to heptamerize and bind lethal factor (LF).  Entry of this complex into the cell is mediated by the endosomal pathway.  Exposure of PA₆₃ to the acidic endosomal environment causes it to undergo a conformational change, resulting in the formation of a 14-stranded β-barrel pore through which LF may gain entry to the cytoplasm.  LF proteolytically modifies the NH₂-termini of members of the MEK family of protein kinases, altering their ability to activate their downstream substrates i.e., extracellular signal regulated protein kinase (ERK), p38 mitogen activated protein kinase (p38 MAPK) and jun kinase (JNK).  In tumor cells, LF activity inhibits the cells’ ability to respond to pro-angiogenic/proliferative stimuli and  causes an inhibition of tumor growth and decreased vascularization.  Image from Singh, Y., Liang, X., and Duesbery, N.S.  2005.  Pathogenesis of Bacillus anthracis; the role of anthrax toxins.  In Microbial Toxins.  T. Proft ed.  Norfolk, U.K.: Horizon Scientific Press, pp. 285–312.