Abstract
Iota toxin is produced by Clostridium perfringens type E strains and associated with diarrhea in cattle and lambs. This binary protein toxin comprises the enzyme component iota a (Ia), which ADP-ribosylates G-actin, and the separate transport component iota b (Ib), which delivers Ia into the cytosol of target cells. Ib binds to cell receptors and forms biologically active toxin complexes with Ia, which cause rounding of adherent cells due to the destruction of the actin cytoskeleton. Here, we report that the human peptide α-defensin-1 protects cultured cells including human colon cells from intoxication with iota toxin. In contrast, the related ß-defensin-1 had no effect, indicating a specific mode of action. The α-defensin-1 did not inhibit ADP-ribosylation of actin by Ia in vitro. Pretreatment of Ib with α-defensin-1 prior to addition of Ia prevented intoxication. Additionally, α-defensin-1 protected cells from cytotoxic effects mediated by Ib in the absence of Ia, implicating that α-defensin-1 interacts with Ib to prevent the formation of biologically active iota toxin on cells. In conclusion, the findings contribute to a better understanding of the functions of α-defensin-1 and suggest that this human peptide might be an attractive starting point to develop novel pharmacological options to treat/prevent diseases associated with iota toxin-producing Clostridium perfringens strains.
INTRODUCTION
Iota toxin is a protein exotoxin produced and released by type E strains of Clostridium perfringens. It belongs to the family of binary actin ADP-ribosylating toxins and represents the prototype of the group of iota-like toxins, which comprises the C. difficile CDT and the C. spiroforme CST as further members. Iota toxin is an enterotoxin associated with sporadic outbreaks of severe diarrhea in livestock, in particular calves and lambs (for review see Stiles 2016). This binary toxin (Stiles and Wilkins 1986) is composed of two separate, non-linked proteins, the enzyme component iota a (Ia) and the transport component iota b (Ib). Both components assemble to form a potent cytotoxin, which is internalized into some mammalian cell types and rapidly induces rounding of adherent cells and disruption of cell monolayers in vitro and exhibits severe toxic effects in vivo such as death of mice. Ia mono-ADP-ribosylates actin at arginine-177 (Vandekerckhove et al.1987; Schering et al.1988), which results in depolymerization of actin filaments, cell-rounding and eventually apoptotic cell death (Hilger et al.2009). Ib binds to its protein receptor LSR (lipolysis-stimulated lipoprotein receptor) on the surface of various cell types (Papatheodorou et al.2011) and likely exploits CD44 as a co-receptor for its cellular uptake (Wigelsworth et al.2012). Ib requires proteolytic activation (Perelle et al.1993; Gibert et al.2000; Blöcker et al.2001) to form heptamers, which represent the biologically active species (Blöcker et al.2001; Nagahama et al.2002, 2004; Stiles et al.2002; Hale et al.2004) and bind Ia (Stiles et al.2000). Following receptor-mediated and clathrin-independent endocytosis via lipid rafts (Stiles et al.2000; Hale et al.2004; Nagahama et al.2004; Gibert et al.2011), the Ib heptamers form pores in the membranes of acidic endosomes and serve as translocation channels for the subsequent transport of Ia from the endosomal lumen into the cytosol (Blöcker et al.2001; Knapp et al.2002, 2015; Gibert et al.2007; Förstner et al.2014). The translocation of Ia across endosomal membranes requires both pH and potential membrane gradients (Gibert et al.2007) and is facilitated by the activities of the host cell chaperones Hsp90 and Hsp70 and of folding helper enzymes like cyclophilins and FK506 binding proteins (Haug, Aktories and Barth 2004; Ernst et al.2015, 2016; Ernst, Schnell and Barth 2017). If the transport of Ia into the cytosol is prevented by pharmacological inhibitors which block the Ib pores or inhibit the activities of the host cell chaperones, cells are protected from intoxication with iota toxin (Nestorovich et al.2011; Förstner et al.2014; Ernst et al.2015, 2016; for review see: Benz and Barth 2017; Ernst, Schnell and Barth 2017). Moreover, in the absence of Ia, Ib alone can exhibit cytotoxic effects (Nagahama et al.2011). Therefore, inhibitors of iota toxin should represent attractive novel means to treat/prevent diseases associated with iota toxin-producing strains of C. perfringens.
Mammalian defensins are small, cysteine-rich cationic peptides with antimicrobial activities (Lehrer, Lichtenstein and Ganz 1993). They play an important role as a first line defense (Ganz 2003) against invading pathogens and are therefore part of the innate immune system (Zasloff 2002; Kim and Kaufmann 2006). In humans, α-defensin is located in neutrophils, which release this peptide at sites of bacterial infection (Fang et al.2003; Kim et al.2006; Ganz 1987; Gudmundsson and Agerberth 1999). During the past decade, it became evident that besides its antibacterial activity, α-defensin-1 can also bind to and neutralize the cytotoxic activity of medically relevant bacterial protein toxins. It was shown for anthrax lethal toxin, a protease (Kim et al.2005) or C. difficile TcdB, a glucosyltransferase (Kim et al.2006; Giesemann, Guttenberg and Aktories 2008) that the defensin inhibits the enzyme activities of these toxins. The activity of TcdB was neutralized by aggregate formation with α-defensin-1 (Giesemann, Guttenberg and Aktories 2008), and it was shown that human defensins facilitate the local unfolding of thermodynamically unstable regions of bacterial protein toxins including TcdA and TcdB in vitro (Kudryashova et al.2014).
Furthermore, it was demonstrated that human α-defensins protect cells from intoxication with diphtheria toxin and Pseudomonas exotoxin A because they neutralize the activities of these bacterial toxins of the mono-ADP-ribosyltransferase family by inhibiting the toxin-catalyzed ADP-ribosylation of the respective cellular substrate molecules (Kim et al.2006). In addition, some, but not all, arginine-specific ADP-ribosylating toxins recognize defensins as a substrate in vitro (Castagnini et al.2012).
Prompted by these results and the tight link between human α-defensin and bacterial ADP-ribosylating toxins, we investigated whether human α-defensin-1 also protects cells from iota toxin. By analyzing the toxin-induced cell rounding, the decrease of the electrical resistance of epithelia, as well as the ADP-ribosylation status of actin as sensitive and specific endpoints to monitor the intoxication process, we found that α-defensin-1 efficiently decreased the cytotoxicity of iota toxin, and investigated the underlying mechanism.
MATERIALS AND METHODS
Protein expression and purification, used inhibitors
The recombinant proteins in this work were expressed and purified as described earlier (Perelle et al.1997). The used α- and β-defensins were purchased from PeptaNova (Sandhausen, Germany)
Cell culture and intoxication experiments
African green monkey kidney (Vero) cells from DSMZ (Braunschweig, Germany) were cultured in minimal essential medium (MEM) containing 10% fetal calf serum (FCS; both GIBCO® life technologies/Thermo Fischer, Darmstadt, Germany), 1 mM sodium pyruvate, 2 mM L-glutamine, 0.1 mM non-essential amino acids and 100 U/ml penicillin/100 μg/ml streptomycin. Human epithelial colorectal adenocarcinoma (CaCo-2) cells (ATCC® HTB-37) were cultured in DMEM including 10% FCS (both GIBCO® life technologies/Thermo Fischer, Darmstadt, Germany), 1 mM sodium-pyruvate, 0.1 mM non-essential amino acids and 100 U/ml penicillin/100 μg/ml streptomycin. For experiments, cells were seeded in 96- or 24-well plates. For intoxication experiments, cells were treated in FCS-free media with the respective compounds and incubated under humidified conditions at 37°C with 5% CO2. After defined time points, pictures were taken using an Axiovert 40CFl microscope from Zeiss (Oberkochen, Germany) connected to a ProgRes C10 CCD camera from Jenoptik (Jena, Germany). Original magnification ×200. Pictures were processed with ImageJ software (NIH). Cell viability was measured by Cell Titer 96® Aqueous One Solution cell proliferation assay (MTS assay, Promega, Mannheim, Germany).
Determination of the half maximal effective concentration of α-defensin-1
Vero cells were intoxicated with iota toxin (0.7 nM Ia + 0.94 nM Ib) and increasing concentrations of α-defensin-1 (0, 0.1, 0.3, 0.6, 1, 3, 6 and 10 μM). After 2 h, the amount of round cells was determined, and the EC50 was calculated using the GraphPad Prism software.
Proteolytic inactivation of α-defensin-1
For in vitro inactivation, α-defensin-1 was treated with trypsin (4.2 μg/mL) for 1 h at 37°C. Thereafter, trypsin was inactivated by incubation with trypsin inhibitor (4.2 μg/mL) for 1 h at 37°C. Afterwards, the digested α-defensin-1 was completed with iota toxin (0.7nM Ia + 0.94 nM Ib) and transferred onto Vero cells. Trypsin and trypsin inhibitor were purchased from Sigma-Aldrich (Deisenhofen, Germany).
In vitro ADP-ribosylation of actin
Cells were grown in 24-well plates. Iota toxin (0.7 nM Ia + 0.94 nM Ib) and either α- or β-defensin-1 was added to the cells. For negative control, cells were treated with medium only. After an incubation period of 6 h, cells were washed with PBS and scraped off in 50 μL ADP-ribosylation-buffer containing 1 mM DTT, 5 mM MgCl2 and 1 mM EDTA, 20 mM Tris-HCl pH 7.5 plus complete protease inhibitor (Roche, Mannheim, Germany). The cells were lysed by freezing to –20°C overnight and subsequent thawing. Equal amounts of lysate protein were incubated with 10 mM biotin-labeled NAD+ (Trevigen/Biozol, Eching, Germany) in the presence of 200 ng freshly added Ia for 30 min at 37°C. By adding SDS sample buffer and subsequent heating at 95°C for 10 min, the enzyme reaction was stopped. After SDS-PAGE, the biotin-labeled, i.e. ADP-ribosylated actin was detected by Western blotting with peroxidase-coupled streptavidin (Sigma Aldrich/Merck, Darmstadt, Germany; 1:2500) and subsequent chemiluminescence reaction. Comparable amounts of blotted protein were confirmed by Western blotting of Hsp90 by a specific antibody (Santa Cruz, Heidelberg, Germany).
Transepithelial electrical resistance measurements
Transepithelial electrical resistance (TEER) measurements were performed using the EVOMX apparatus (WPI, Berlin, Germany) provided with the STX2 electrode. To get a confluent monolayer, 0.7 × 105 CaCo-2 cells were seeded on 24-well Hanging Inserts (Millicell® Cell Culture Inserts, Merck, Darmstadt, Germany) and incubated for 3 days at 37°C and with 5% CO2. For TEER measurements, iota toxin was added apically and cells were incubated with the toxin and the defensins in complete medium. Raw data were transformed into unit area resistance by subtraction of the blank and by multiplying the resulting data by 0.3 cm2.
In vitro pre-incubation of iota-toxin with α- or β-defensin-1
Ia (0.7 nM) plus Ib (0.94 nM) or the individual single components Ia and Ib were incubated in vitro on a heating block at 37°C, 400 rpm for 30 min in the presence or absence of α- or β-defensin-1 (each 6 μM). Afterwards, the respective complementary component of iota toxin was freshly supplemented and samples were transferred onto Vero cells seeded in a 96-well plate. Cells were incubated at 37°C with 5% CO2 for 5 h. The toxin-induced changes in cell morphology were recorded and percentages of round cells determined from the pictures.
Detection of cell-associated Ib or Ia
For detection of cell-associated Ib, Vero cells were incubated at 37°C with Ib (9.4 nM) in the presence or absence of α-defensin-1 (6 μM). After 20 min, the cells were extensively washed with PBS, lysed, scraped off in hot SDS-sample buffer and transferred to SDS-PAGE with subsequent Western blotting. Bound or cell-associated Ib was detected using an Ib-antibody (1:1000, Perelle et al.1993) in combination with a HRP-coupled secondary antibody (SantaCruz, Heidelberg, Germany, goat anti-rabbit IgG-HRP, 1:2500). Comparable amounts of loaded protein were confirmed by Western blotting of GAPDH by a specific antibody (SantaCruz, Heidelberg, Germany). For detection of cell-associated Ia, Vero cells were incubated at 37°C with iota toxin (7.3 nM Ia + 9.4 nM Ib) and treated as described above. Cell-associated Ia was detected using an Ia-antibody (1:3000, Perelle et al.1993) in combination with a HRP-coupled secondary antibody (SantaCruz, Heidelberg, Germany, goat anti-rabbit IgG-HRP, 1:2500). Comparable amounts of loaded protein were confirmed by Western blotting of Hsp90.
Reproducibility of the experiments
All experiments were performed independently for at least two times. For each cell experiment, two biological independent experiments in triplicate were performed. Results from representative experiments are shown in the figures.
RESULTS AND DISCUSSION
The presence of human α-defensin-1 leads to delayed intoxication of cells with iota toxin
Prompted by earlier observations that human α-defensin-1 protects cultured cells from intoxication with bacterial exotoxins, we tested the effect of α-defensin-1 and ß-defensin-1 on the intoxication of African Green Monkey kidney (Vero) cells with iota toxin. This is an established mammalian cell line to investigate the effects of iota toxin because Vero cells are highly sensitive towards this toxin and the intoxication process can be easily followed by monitoring the toxin-induced cell rounding, which is a highly sensitive and specific endpoint to monitor the cytotoxic effect of iota toxin (Blöcker et al.2001). Vero cells were incubated with a mixture of Ia plus Ib in the medium in the absence or presence of each defensin and the amount of round cells was determined over time. For control, cells were left untreated. As shown in Fig. 1A, the percentage of round cells was increasing over time when cells were incubated with iota toxin alone. Indeed, in the presence of α-defensin-1, there was less cell rounding, indicating that α-defensin-1 was able to significantly delay the intoxication with iota toxin. In contrast, treatment with the ß-defensin-1 had no effect on the amount of round cells, indicating that this related defensin did not inhibit the intoxication of Vero cells with iota toxin. This result suggests that the inhibition of iota toxin by α-defensin-1 occurs via a specific mechanism because the inhibitory effect of α-defensin-1 was reduced after digestion with trypsin (Fig. 1B). Noteworthy, incubation of the cells with the defensins, even for longer periods, had no effect on cell morphology, implicating that these human peptides have no adverse effects on the cells. Next, the concentration-dependent effect of α-defensin-1 was investigated. Vero cells were incubated with iota toxin and increasing concentrations of α-defensin-1 (ranging from 0 to 10 μM). A significant reduction of the amount of rounded Vero cells at a defined time point (2 h) was observed for α-defensin-1-concentrations between 0.6 and 10 μM compared to cells without α-defensin-1 (Fig. 1C). The half-maximal effective concentration of α-defensin-1 under these particular conditions was 3 μM. Furthermore, comparable results were obtained when human colon carcinoma cells (CaCo-2) were used instead of Vero cells (Fig. 1D). Although the intoxication of CaCo-2 cells was faster compared to Vero cells, there was a significant delay in intoxication caused by α-defensin-1, too. This indicates that the observed inhibitory effect of α-defensin-1 is not restricted to a certain cell line and that also human cells are protected from iota toxin by α-defensin-1 but not by ß-defensin-1, which might be of more medically relevance for other binary actin-ADP-ribosylating toxins of this family, such as C. difficile CDT.
α-defensin-1 delayed intoxication of Vero and CaCo-2 cells with iota toxin. (A) Vero cells were incubated for 4 h with iota toxin (0.7 nM Ia + 0.94 nM Ib) together with either α-defensin-1 (6 μM) or β-defensin-1 (6 μM). For control, cells were left untreated. The time-dependent inhibition of iota toxin by α-defensin-1 is shown. Values are given as mean ± SD (n = 3). Shown is a typical result from two independent experiments (each experiment performed in triplicate). Scale bar: 50 μm. (B) α-defensin-1 was incubated with trypsin for 1 h at 37°C and subsequent incubation with trypsin inhibitor for 1 h at 37°C. Then, the approach was completed with iota toxin (0.7 nM Ia + 0.94 nM Ib) and transferred onto Vero cells. After 3 h, the amount of rounded cells was determined. The experiment was performed two times in triplicate. Values are given as mean ± SD (n = 3). (C) Half maximal effective concentration of α-defensin-1 after 2 h was estimated by intoxicating Vero cells with iota toxin (0.7 nM Ia + 0.94 nM Ib) in the presence of different concentrations of α-defensin-1. Values are given as mean ± SD (n = 3) from two independent experiments performed in triplicate. (D) CaCo-2 cells were incubated with iota toxin (0.7 nM Ia + 0.94 nM Ib) in the presence of either α-defensin-1 or β-defensin-1 (each 6 μM). For control, cells were left untreated. Representative images from two independent experiments after 2 h are shown. Each experiment was designed in triplicate. Significance was tested by using Student's t test. Values refer to samples treated with iota toxin (ns = not significant, *P < 0.05, **P < 0.01, ***P < 0.001). Scale bar: 50 μm.
Next, the ADP-ribosylation status of actin from Vero cells was analyzed to investigate another highly specific endpoint of the intoxication of cells by iota toxin. To this end, the cells were treated as described before. After the incubation period, the cells were lysed and equal amounts of lysate protein were subjected to in vitro ADP-ribosylation with fresh Ia as enzyme and biotin-labeled NAD+ as co-substrate. The biotin-labeled, i.e. ADP-ribosylated actin was then analyzed by Western blotting (Fig. 2). The protective effect of α-defensin-1 became evident, while also in this approach ß-defensin-1 had no inhibitory effect towards iota toxin. Here, a strong signal of biotin-labeled, i.e. ADP-ribosylated actin indicates that in the living cells the actin was not modified. In contrast, a weak signal in the Western blot indicates that most of the actin was already ADP-ribosylated by iota toxin in the cytosol of the living cells during the incubation with the toxin, and therefore is no more a substrate in the subsequent in vitro ADP-ribosylation reaction (Blöcker et al.2001).
Incubation of Vero cells with α-defensin-1 protects intracellular actin from iota toxin-catalyzed ADP-ribosylation. Vero cells were incubated with iota toxin (0.7 nM Ia + 0.94 nM Ib) in combination with either α-defensin-1 or β-defensin-1 (each 6 μM). After 6 h, cells were lysed and the ADP-ribosylation status of their actin analyzed by in vitro ADP-ribosylation and Western blotting (lower panel). Intensity of signals was quantified by densitometry. Depicted values display the percentage of the in vitro ADP-ribosylated actin of untreated cells and are normalized on the respective amount of Hsp90. Values are given as mean ± SEM (n = 3). Significance was tested by using Student's t test. Values refer to the sample treated with toxin alone (ns = not significant, *P < 0.05). Shown is a representative Western blot from two independent experiments performed in duplicate.
Furthermore, the effect of α-defensin-1 and ß-defensin-1 on the intoxication of CaCo-2 cells with iota toxin was investigated by analyzing the TEER of confluently grown cells, which mimics the situation of an epithelium in vitro. The result shown in Fig. 3A clearly indicates the delay in intoxication when α-defensin-1, but not ß-defensin-1, was present. The integrity of the CaCo-2 epithelium after treatment with iota toxin remained more intact in the presence of α-defensin-1, thus confirming the observations obtained by the morphological and biochemical analysis described before. Also in this approach, the defensins alone had no effect on the integrity of the epithelium formed by CaCo-2 cells in vitro (Fig. 3B).
α-defensin-1 but not β-defensin-1 protects the epithelial integrity of CaCo-2 monolayers from intoxication with iota toxin. (A) TEER measurements were performed using CaCo-2 cells with iota toxin (0.7 nM Ia + 0.94 nM Ib) in the presence or absence of either α-defensin-1 or β-defensin-1 (each 6 μM). Shown are the results from two independent experiments performed at least in duplicate. The values are given as mean ± SD (n = 5). (B) CaCo-2 monolayer was treated with the respective defensin (each 6μM) in the absence of toxin and TEER measurements were performed. Shown is the result from two experiments each performed in duplicate and values are given as mean ± SD (n = 4). Significance was tested by using Student's t test. Values refer to untreated cells (ns = not significant, *P < 0.05, **P < 0.01).
Incubation with α-defensin-1 neither inhibits the enzyme activity of Ia nor the binding of the iota toxin to cells
Next, we investigated the molecular mechanism underlying the protection of the intoxication of cells with iota toxin by α-defensin-1 in more detail. Prompted by earlier results obtained for other ADP-ribosylating toxins (Kim et al.2006), the effect of this peptide on the enzyme activity of Ia was analyzed by in vitro ADP-ribosylation of actin. For this purpose, Vero or CaCo-2 lysate was used as a source for actin and incubated at 37°C with Ia and biotin-NAD+ and the biotin-labeled, i.e. ADP-ribosylated actin was determined by Western blotting (Fig. 4). Taken together, neither α-defensin-1 nor ß-defensin-1 obviously decreased the amount of ADP-ribosylated actin, suggesting that the defensins did not interfere with the ADP-ribosyltransferase activity of Ia under these experimental conditions. This result is in some contrast to the earlier findings that this defensin inhibits in vitro the enzyme activities of other bacterial toxins like diphtheria toxin (Kim et al.2006), anthrax lethal toxin (Kim et al.2005) or TcdB (Giesemann, Guttenberg and Aktories2008) and suggests a different, specific mode of action between human α-defensin-1 and iota toxin.
The defensins do not inhibit the ADP-ribosylation of actin by Ia in vitro. Vero lysate as well as CaCo-2 lysate was incubated with Ia (200 ng) in the presence or absence of either α-defensin-1 or β-defensin-1 (6 μM) at 37°C for 30 min. After heat-denaturation and subsequent SDS-PAGE, the biotin-labeled, i.e. ADP-ribosylated actin was detected by Western blotting. Comparable amounts of blotted cell lysate proteins were confirmed by Western blotting with an antibody against Hsp90 (upper panel). Intensity of signals was quantified by densitometry. Depicted values display the percentage of ADP-ribosylated actin (lower panel) and are normalized on the amount of Hsp90. The experiment was performed two times with comparable results. Values are given as mean ± SEM (n = 3). A typical Western blot result is shown. Significance was tested by using Student's t test. Values refer to the signal for ADP-ribosylated actin in the Ia treatment sample (ns = not significant).
Prompted by the finding of several other groups that α-defensin-1 binds to various bacterial protein toxins, which most likely induces conformational and functional changes of the toxins (Kim et al.2005, 2006; Giesemann, Guttenberg and Aktories 2008; Kudryashova et al.2014), we tested whether pre-incubation of the single components of iota toxin with this peptide also neutralize its biological activity. Therefore, either Ia alone or Ib alone were incubated in vitro with α-defensin-1 or β-defensin-1 and then the mixtures were applied to the cells, Ia in combination with fresh Ib and Ib in combination with fresh Ia. For control, Ia or Ib were pre-incubated in the absence of defensins and applied to cells as described before. Cells were incubated at 37°C and the amount of round cells was determined after various periods. As shown in Fig. 5, pre-incubation of each toxin component alone with α-defensin-1, but not with ß-defensin-1 reduced the amount of round cells, clearly indicating the protective effect of α-defensin-1. Pre-incubation of Ia with α-defensin-1 prior to application to cells resulted in reduced activity of iota toxin (Fig. 5A). Nevertheless, the inhibitory effect of α-defensin-1 was most obvious when it was pre-incubated with Ib (Fig. 5B), suggesting that the biological activity of Ib was diminished by α-defensin-1. However, pre-incubation of Ib at 37°C decreased its biological activity to some extent.
Pre-incubation of iota toxin or single iota-toxin components with α-defensin-1 protects Vero cells from intoxication. (A) The single component Ia (0.7 nM) was pre-incubated in vitro with either α- or β-defensin-1 (each 6 μM) or without defensins for control. Pre-incubation took place on a heating block at 37°C and 400 rpm. After 30 min, Ia was completed with fresh Ib (0.94 nM). (B) Ib (0.94 nM) was pre-incubated at 37°C with either α- or β-defensin-1 (each 6 μM). After 30 min, Ib was completed with fresh Ia (0.7 nM). Both approaches were transferred onto Vero cells and incubated at 37°C. The same control was used as in (A). After 5 h, pictures were taken and the amount of round cells was determined. Representative images from two independent experiments performed in triplicate are displayed. Values are given as mean ± SD (n = 3) and refer to the respective positive control (ns = not significant, ***P < 0.001). Scale bar: 50 μm.
Prompted by these findings, we tested, whether α-defensin-1 was able to prevent cells from the cytotoxic effects mediated by Ib in the absence of Ia. It was reported earlier that the transport component Ib alone is able to provoke cytotoxic effects by inducing rapid cell necrosis (Nagahama et al.2011). On the basis of these findings, we incubated Vero cells with a concentration of Ib that is known to induce cytotoxicity, added either α- or β-defensin-1 and monitored the effect of Ib on the cells in terms of changes in cell morphology and the amount of viable cells. Indeed, there was a clear protective effect of α-defensin-1, but not of ß-defensin-1, towards the effects mediated by Ib (Fig. 6A). Furthermore, cells, which were treated with Ib in the presence of α-defensin-1, showed comparable viability to untreated cells while β-defensin-1 was not able to rescue cells from the intoxication with Ib (Fig. 6B). Finally, it was investigated whether α-defensin-1 reduced the binding of Ib to cells or the binding of Ia to Ib. First, Vero cells were incubated with Ib in the presence or absence of α-defensin-1. After 20 min, the cells were washed, lysed and cell-associated Ib was detected by Western blotting with a specific antibody against Ib. There was no obvious difference in the amount of Ib in the presence or absence of α-defensin-1, suggesting that receptor binding of Ib is not affected by α-defensin-1 (Fig. 6C). Next, Vero cells were incubated with complete iota toxin in the presence or absence of α-defensin-1, treated as described above and cell-associated Ia was detected by Western blotting with a specific antibody. The amount of Ia was widely comparable, regardless whether α-defensin-1 was present or not (Fig. 6D).
α-defensin-1 protects Vero cells from intoxication with Ib but does not inhibit binding of Ib to cells. (A) Vero cells were incubated with Ib (2.7 nM) in the presence or absence of either α- or β-defensin-1 (each 6 μM). After 1 h, representative images are shown. Scale bar: 50 μm. A typical result from two independent experiments is shown (each experiment was performed in triplicate). (B) After a total incubation time of 3 h, the amount of viable cells was measured by MTS assay. Values are given as mean ± SD (n = 6). Significance was tested by using Student's t test. Values refer to untreated negative control (ns = not significant, ***P < 0.001). (C) α-defensin-1 does not inhibit binding of Ib to cells. Vero cells were treated with Ib (9.4 nM) in the presence of α-defensin-1 (6 μM) but in the absence of Ia. After 20 min of incubation, the amount of cell-associated Ib was detected. Depicted values display the percentage of Ib and are normalized on the amount of GAPDH (n = 6). Shown is a typical result from two independent experiments performed in triplicate. Significance was tested by using Student's t test. Values refer to Ib (ns = not significant). (D) α-defensin-1 does not inhibit binding of Ia to Ib. Vero cells were treated with iota toxin (7.3 nM Ia + 9.4 nM Ib) for 20 min in the presence or absence of α-defensin-1. Then, the amount of cell-associated Ia was detected. The picture shows a typical Western blot result. The experiment was performed two times in triplicate. Values refer to Ia and are normalized on the amount of Hsp90 (n = 5). Significance was tested by using Student's t test (ns = not significant).
Taken together, these results suggest that α-defensin-1 inhibits the mode of action of iota toxin via an interaction with Ib. However, the underlying molecular mechanism remains speculative. If α-defensin-1 influences the structure of Ib, this would not result in obvious consequences for binding of Ib to its cellular receptor, nor for binding of Ia to Ib. It is also conceivable that α-defensin-1 influences the ability of Ib to form transmembrane pores. Thereby, α-defensin-1 could act as a pore blocker preventing translocation of Ia through Ib-formed pores when Ib is applied together with Ia without influencing the ability of Ia to interact with Ib. Nevertheless, neutralization of iota toxin is not only based on unspecific interactions between the toxin and a positively charged peptide since in all experiments no comparable results were observed with the related β-defensin-1, which shares the same features as α-defensin-1 (i.e. amount of disulfide bonds, cationic charge). It will be important to test whether α-defensin-1 also binds to and neutralizes the further members of the family of binary actin ADP-ribosylating toxins, which share significant structural and functional similarities with iota toxin. In this context, we obtained first evidence that the cytotoxicity of the medically relevant binary CDT from hypervirulent strains of C. difficile can be neutralized by human α-defensin-1 (Fischer et al. in preparation). CDT contributes to more severe forms of C. difficile-associated diseases including pseudomembranous colitis and is closely related to iota toxin (for review see Aktories, Schwan and Jank 2017). Its transport component shows 82% sequence homology to Ib (for review see Stiles 2016) and exploits the same cellular receptor LSR (Papatheodorou et al.2011) and used CD44 for its cellular uptake (Wigelsworth et al.2012), which is comparable to the uptake of iota toxin (Blöcker et al.2001; Haug, Aktories and Barth 2004; Ernst et al.2015, 2016, 2017). In this context, human patients with bacterial infections showed dramatically increased blood concentrations of HNP1-3. Previous studies have shown that the concentrations of HNP1-3 were 3.2 times the value for healthy patients, resulting in a blood concentration of around 35 ng/μL (approx. 10 μM) (Ihi et al.1997). When neutrophils ingest bacteria, intracellular granules release α-defensins with local concentrations up to 10 mg/mL (approx. 3 mM) onto the pathogen (Ganz 1987). However, plasma level of HNPs from patients with sepsis or bacterial meningitis can reach concentrations up to 100 mg/mL (approx. 29 mM) (Suarez-Carmona et al.2015). Taken together, our results obtained for iota toxin as a prototype of clostridial binary actin ADP-ribosylating toxins extends the understanding that human α-defensin specifically combats medically relevant bacterial protein toxins. The application of toxin-neutralizing human α-defensin might be an attractive therapeutic option against infections with toxin-producing bacteria, in particular if such bacteria are (multi) resistant towards the established antibiotics, in particular because of their human origin, no severe side effects should be expected after application of such peptides.
ACKNOWLEDGMENTS
Stephan Fischer is a member of the International Graduate School in Molecular Medicine Ulm (IGradU) and thanks IGradU for the support.
FUNDING
The work was supported by the: German Research Foundation, SFB 1279 (project C02).
Conflict of interest. None declared.
