Chloroquine Analog Interaction with C2- and Iota-Toxin in Vitro and in Living Cells

Chloroquine-analogues block anthrax protective antigen channels in steady-state and kinetic studies

The tripartite anthrax toxin from Bacillus anthracis represents the prototype of A-B type of toxins, where the effector A (an enzymatic subunit) is transported with the help of a binding component B into a target cell. Anthrax toxin consists of three different molecules, two effectors, lethal factor (LF) and edema factor (EF) and the binding component also known as protective antigen (PA). PA forms heptamers or octamers following binding to host cell's receptors and mediates the translocation of the effectors into the cytosol via the endosomal pathway. The cation-selective PA₆₃-channel is able to reconstitute in lipid membranes and can be blocked by chloroquine and other heterocyclic compounds. This suggests that the PA₆₃-channel contains a binding site for quinolines. In this study, we investigated the structure-function relationship of different quinolines for the block of the PA₆₃-channel. The affinity of the different chloroquine analogues to the PA₆₃-channel as provided by the equilibrium dissociation constant was measured using titrations. Some quinolines had a much higher affinity to the PA₆₃-channel than chloroquine itself. We also performed ligand-induced current noise measurements using fast Fourier transformation to get insight in the kinetics of the binding of some quinolines to the PA₆₃-channel. The on-rate constants of ligand binding were around 10⁸M^-1·s^-1 at 150mM KCl and were only little dependent on the individual quinoline. The off-rates varied between 4s^-1 and 160s^-1 and depended much more on the structure of the molecules than the on-rate constants. The possible use of the 4-aminoquinolines as a therapy is discussed.

The Pore-Forming Subunit C2IIa of the Binary Clostridium botulinum C2 Toxin Reduces the Chemotactic Translocation of Human Polymorphonuclear Leukocytes

The binary C2 toxin of Clostridium (C.) botulinum consists of two non-linked proteins, the enzyme subunit C2I and the separate binding/transport subunit C2II. To exhibit toxic effects on mammalian cells, proteolytically activated C2II (C2IIa) forms barrel-shaped heptamers that bind to carbohydrate receptors which are present on all mammalian cell types. C2I binds to C2IIa and the toxin complexes are internalized via receptor-mediated endocytosis. In acidified endosomal vesicles, C2IIa heptamers change their conformation and insert as pores into endosomal membranes. These pores serve as translocation-channels for the subsequent transport of C2I from the endosomal lumen into the cytosol. There, C2I mono-ADP-ribosylates G-actin, which results in depolymerization of F-actin and cell rounding. Noteworthy, so far morphological changes in cells were only observed after incubation with the complete C2 toxin, i.e., C2IIa plus C2I, but not with the single subunits. Unexpectedly, we observed that the non-catalytic transport subunit C2IIa (but not C2II) alone induced morphological changes and actin alterations in primary human polymorphonuclear leukocytes (PMNs, alias neutrophils) from healthy donors ex vivo, but not macrophages, epithelial and endothelial cells, as detected by phase contrast microscopy and fluorescent microscopy of the actin cytoskeleton. This suggests a PMN selective mode of action for C2IIa. The cytotoxicity of C2IIa on PMNs was prevented by C2IIa pore blockers and treatment with C2IIa (but not C2II) rapidly induced Ca²⁺ influx in PMNs, suggesting that pore-formation by C2IIa in cell membranes of PMNs is crucial for this effect. In addition, incubation of primary human PMNs with C2IIa decreased their chemotaxis ex vivo through porous culture inserts and in co-culture with human endothelial cells which is closer to the physiological extravasation process. In conclusion, the results suggest that C2IIa is a PMN-selective inhibitor of chemotaxis. This provides new knowledge for a pathophysiological role of C2 toxin as a modulator of innate immune cells and makes C2IIa an attractive candidate for the development of novel pharmacological strategies to selectively down-modulate the excessive and detrimental PMN recruitment into organs after traumatic injuries.

Characterization and Pharmacological Inhibition of the Pore-Forming Clostridioides difficile CDTb Toxin

The clinically highly relevant Clostridioides (C.) difficile releases several AB-type toxins that cause diseases such as diarrhea and pseudomembranous colitis. In addition to the main virulence factors Rho/Ras-glycosylating toxins TcdA and TcdB, hypervirulent strains produce the binary AB-type toxin CDT. CDT consists of two separate proteins. The binding/translocation B-component CDTb facilitates uptake and translocation of the enzyme A-component CDTa to the cytosol of cells. Here, CDTa ADP-ribosylates G-actin, resulting in depolymerization of the actin cytoskeleton. We previously showed that CDTb exhibits cytotoxicity in the absence of CDTa, which is most likely due to pore formation in the cytoplasmic membrane. Here, we further investigated this cytotoxic effect and showed that CDTb impairs CaCo-2 cell viability and leads to redistribution of F-actin without affecting tubulin structures. CDTb was detected at the cytoplasmic membrane in addition to its endosomal localization if CDTb was applied alone. Chloroquine and several of its derivatives, which were previously identified as toxin pore blockers, inhibited intoxication of Vero, HCT116, and CaCo-2 cells by CDTb and CDTb pores in vitro. These results further strengthen pore formation by CDTb in the cytoplasmic membrane as the underlying cytotoxic mechanism and identify pharmacological pore blockers as potent inhibitors of cytotoxicity induced by CDTb and CDTa plus CDTb.

The cytotoxic effect of Clostridioides difficile pore-forming toxin CDTb

Clostridioides (C.) difficile is clinically highly relevant and produces several AB-type protein toxins, which are the causative agents for C. difficile-associated diarrhea and pseudomembranous colitis. Treatment with antibiotics can lead to C. difficile overgrowth in the gut of patients due to the disturbed microbiota. C. difficile releases large Rho/Ras-GTPase glucosylating toxins TcdA and TcdB, which are considered as the major virulence factors for C. difficile-associated diseases. In addition to TcdA and TcdB, C. difficile strains isolated from severe cases of colitis produce a third toxin called CDT. CDT is a member of the family of clostridial binary actin ADP-ribosylating toxins and consists of two separate protein components. The B-component, CDTb, binds to the receptor and forms a complex with and facilitates transport and translocation of the enzymatically active A-component, CDTa, into the cytosol of target cells by forming trans-membrane pores through which CDTa translocates. In the cytosol, CDTa ADP-ribosylates G-actin causing depolymerization of the actin cytoskeleton and, eventually, cell death. In the present study, we report that CDTb exhibits a cytotoxic effect in the absence of CDTa. We show that CDTb causes cell rounding and impairs cell viability and the epithelial integrity of CaCo-2 monolayers in the absence of CDTa. CDTb-induced cell rounding depended on the presence of LSR, the specific cellular receptor of CDT. The isolated receptor-binding domain of CDTb was not sufficient to cause cell rounding. CDTb-induced cell rounding was inhibited by enzymatically inactive CDTa or a pore-blocker, implying that CDTb pores in cytoplasmic membranes contribute to cytotoxicity.

Human peptide α-defensin-1 interferes with Clostridioides difficile toxins TcdA, TcdB, and CDT

The human pathogenic bacterium Clostridioides difficile produces two exotoxins TcdA and TcdB, which inactivate Rho GTPases thereby causing C. difficile-associated diseases (CDAD) including life-threatening pseudomembranous colitis. Hypervirulent strains produce additionally the binary actin ADP-ribosylating toxin CDT. These strains are hallmarked by more severe forms of CDAD and increased frequency and severity. Once in the cytosol, the toxins act as enzymes resulting in the typical clinical symptoms. Therefore, targeting and inactivation of the released toxins are of peculiar interest. Prompted by earlier findings that human α-defensin-1 neutralizes TcdB, we investigated the effects of the defensin on all three C. difficile toxins. Inhibition of TcdA, TcdB, and CDT was demonstrated by analyzing toxin-induced changes in cell morphology, substrate modification, and decrease in transepithelial electrical resistance. Application of α-defensin-1 protected cells and human intestinal organoids from the cytotoxic effects of TcdA, TcdB, CDT, and their combination which is attributed to a direct interaction between the toxins and α-defensin-1. In mice, the application of α-defensin-1 reduced the TcdA-induced damage of intestinal loops in vivo. In conclusion, human α-defensin-1 is a specific and potent inhibitor of the C. difficile toxins and a promising agent to develop novel therapeutic options against C. difficile infections.

Inhibition of Pore-Forming Proteins

Perforation of cellular membranes by pore-forming proteins can affect cell physiology, tissue integrity, or immune response. Since many pore-forming proteins are toxins or highly potent virulence factors, they represent an attractive target for the development of molecules that neutralize their actions with high efficacy. There has been an assortment of inhibitors developed to specifically obstruct the activity of pore-forming proteins, in addition to vaccination and antibiotics that serve as a plausible treatment for the majority of diseases caused by bacterial infections. Here we review a wide range of potential inhibitors that can specifically and effectively block the activity of pore-forming proteins, from small molecules to more specific macromolecular systems, such as synthetic nanoparticles, antibodies, antibody mimetics, polyvalent inhibitors, and dominant negative mutants. We discuss their mechanism of inhibition, as well as advantages and disadvantages.

Exploring the Nature of Cationic Blocker Recognition by the Anthrax Toxin Channel

Obstructing conductive pathways of the channel-forming toxins with targeted blockers is a promising drug design approach. Nearly all tested positively charged ligands have been shown to reversibly block the cation-selective channel-forming protective antigen (PA₆₃) component of the binary anthrax toxin. The cationic ligands with more hydrophobic surfaces, particularly those carrying aromatic moieties, inhibited PA₆₃ more effectively. To understand the physical basis of PA₆₃ selectivity for a particular ligand, detailed information is required on how the blocker structural elements (e.g., positively charged and aromatic groups) influence the molecular kinetics of the blocker/channel binding reactions. In this study, we address this problem using the high-resolution single-channel planar lipid bilayer technique. Several structurally distinct cationic blockers, namely per-6-S-(3-amino) propyl-β-cyclodextrin, per-6-S-(3-aminomethyl) benzyl-α-cyclodextrin, per-6-S-(3-aminomethyl) benzyl-β-cyclodextrin, per-6-S-(3-aminomethyl) benzyl-γ-cyclodextrin, methyltriphenylphosphonium ion, and G0 polyamidoamine dendrimer are tested for their ability to inhibit the heptameric and octameric PA₆₃ variants and PA₆₃F427A mutant. The F427 residues form a hydrophobic constriction region inside the channel, known as the "ϕ-clamp." We show that the cationic blockers interact with PA₆₃ through a combination of forces. Analysis of the binding reaction kinetics suggests the involvement of cation-π, Coulomb, and salt-concentration-independent π-π or hydrophobic interactions in the cationic cyclodextrin binding. It is possible that these blockers bind to the ϕ-clamp and are also stabilized by the Coulomb interactions of their terminal amino groups with the water-exposed negatively charged channel residues. In PA₆₃F427A, only the suggested Coulomb component of the cyclodextrin interaction remains. Methyltriphenylphosphonium ion and G0 polyamidoamine dendrimer, despite being positively charged, interact primarily with the ϕ-clamp. We also show that seven- and eightfold symmetric cyclodextrins effectively block the heptameric and octameric forms of PA₆₃ interchangeably, adding flexibility to the earlier formulated blocker/target symmetry match requirement.

Impact of Dendrimer Terminal Group Chemistry on Blockage of the Anthrax Toxin Channel: A Single Molecule Study

Nearly all the cationic molecules tested so far have been shown to reversibly block K⁺ current through the cation-selective PA₆₃ channels of anthrax toxin in a wide nM-mM range of effective concentrations. A significant increase in channel-blocking activity of the cationic compounds was achieved when multiple copies of positively charged ligands were covalently linked to multivalent scaffolds, such as cyclodextrins and dendrimers. Even though multivalent binding can be strong when the individual bonds are relatively weak, for drug discovery purposes we often strive to design multivalent compounds with high individual functional group affinity toward the respective binding site on a multivalent target. Keeping this requirement in mind, here we perform a single-channel/single-molecule study to investigate kinetic parameters of anthrax toxin PA₆₃ channel blockage by second-generation (G2) poly(amido amine) (PAMAM) dendrimers functionalized with different surface ligands, including G2-NH₂, G2-OH, G2-succinamate, and G2-COONa. We found that the previously reported difference in IC₅₀ values of the G2-OH/PA₆₃ and G2-NH₂/PA₆₃ binding was determined by both on- and off-rates of the reversible dendrimer/channel binding reaction. In 1 M KCl, we observed a decrease of about three folds in k o n and a decrease of only about ten times in t r e s with G2-OH compared to G2-NH₂. At the same time for both blockers, k o n and t r e s increased dramatically with transmembrane voltage increase. PAMAM dendrimers functionalized with negatively charged succinamate, but not carboxyl surface groups, still had some residual activity in inhibiting the anthrax toxin channels. At 100 mV, the on-rate of the G2-succinamate binding was comparable with that of G2-OH but showed weaker voltage dependence when compared to G2-OH and G2-NH₂. The residence time of G2-succinamate in the channel exhibited opposite voltage dependence compared to G2-OH and G2-NH₂, increasing with the cis-negative voltage increase. We also describe kinetics of the PA₆₃ ion current modulation by two different types of the "imperfect" PAMAM dendrimers, the mixed-surface G2 75% OH 25% NH₂ dendrimer and G3-NH₂ dendron. At low voltages, both "imperfect" dendrimers show similar rate constants but significantly weaker voltage sensitivity when compared with the intact G2-NH₂ PAMAM dendrimer.