TcdA1 of Photorhabdus luminescens: Electrophysiological Analysis of Pore Formation and Effector Binding

Abstract

Tc toxins are widely distributed among different gram-negative and gram-positive bacteria, where they act as pathogenicity factors. The toxins are composed of different components that form oligomers for biological activity. Lipid bilayer experiments were performed with the TcdA1 component of the Tc toxin from Photorhabdus luminescens, which preferentially kills insects by actin polymerization. TcdA1 was able to increase the specific conductance of artificial lipid bilayer membranes by the formation of ion-permeable channels. The channels had on average a single-channel conductance of 125 pS in 150 mM KCl and were found to be cation selective. The single-channel conductance of the TcdA1-channels was only moderately dependent on the bulk aqueous KCl concentration, which indicated point-charge effects on the channel properties. Experiments to study the voltage dependence of the TcdA1 channel demonstrated that it is reconstituted in a fully oriented way when it is added to only one side of the lipid bilayer membrane. A combination of biologically active components (TccC3) and a possible chaperone (TcdB2) blocked the TcdA1-mediated conductance efficiently in a dose-dependent manner when they were added to the cis side of the membrane. The half-saturation constant for binding of TcdB2-TccC3 to TcdA1 is in the low nanomolar range.

Introduction

Several intracellularly acting bacterial protein toxins that are known to enter cells by endocytosis have been shown to produce pores. These include diphtheria toxin (1–3), neurotoxins (4–6), and the binding components of anthrax toxin, Clostridium botulinum C2 and C. perfringens iota toxins (7–10). Studies of these toxins (1,10–12) have provided some evidence that toxin translocation and channel formation are related phenomena.

Tc toxins are produced by entomopathogenic Photorhabdus luminescens, which exists in symbiosis with nematodes. After insect larvae are invaded by nematodes, the bacteria are released and start to produce toxins that eventually kill the insect, thereby generating a food resource for bacteria and nematodes. Tc toxins consist of three separate components: TcA, TcB, and TcC. Whereas TcA is suggested to be involved in membrane receptor binding and translocation, the TcC component possesses biological activity inside the target cells for which they need to translocate through the cell membranes, presumably via the endosomal pathway (13). Recently, it was shown that the Tc toxins TccC3 and TccC5 ADP-ribosylate actin and RhoA, respectively, thereby inducing actin clustering (13). The function of TcB is still poorly understood, but it may have a role as a chaperone. Moreover, several homologs of each Tc component exist, which complicates the toxin repertoire of the bacteria.

Recently, the structure of the TcdA1 oligomer was studied by cryo-electron microscopy (cryo-EM) and single-particle analysis, resulting in 6.3 Å resolution. The resultant data suggest that TcdA1 forms a syringe-like pentameric structure with a molecular mass of ∼1.41 MDa and a funnel-like central pore inside the pentamer that conducts ions (14). Whereas many pore-forming toxins, including Staphylococcus aureus α-hemolysin (15), Aeromonas hydrophila aerolysin (16), cholesterol-dependent cytolysins (17), and the binding components of anthrax toxin (10) and C. botulinum C2 toxin exhibit β-barrel structures (18), the structure of TcdA1 is formed mainly by α-helices. The high content of α-helices of the TcdA1 pentamer suggests that the Tc toxin system has a completely different structure compared with the delivery systems of anthrax toxin, C. botulinum C2 toxin, and C. perfringens iota toxins (10,14,19,20).

Here, we studied the influence of the TcA component TcdA1 and other components of the toxin complex on the conductivity of artificial lipid bilayer membranes. We report that TcdA1, but not other components of the toxin, such as TcdB2 and TccC3, is able to induce the formation of small ion-selective channels in artificial lipid bilayer membranes. The channels are cation selective due to negative point charges at the channel mouth. Furthermore, the results suggest that the TcdA1 channel has a defined orientation when the protein is added to only one side of the membrane. The TcdA1 channel also appears to be asymmetric with respect to voltage dependence and TcdB2-TccC3 binding. The titration experiments with TcdB2-TccC3 provide interesting insight into the molecular requirement for effector binding of the TcdA1 channel.

Material and Methods

Materials

All salts were obtained from Merck (Darmstadt, Germany; analytical grade). Ultrapure water was obtained by passing deionized water through Milli-Q equipment (Millipore, Bedford, MA). Aqueous salt solutions were prepared in ultrapure water. Then 10 mM MES-KOH was added to the salt solutions and their pH was adjusted to pH 6. DNase I and Protease Inhibitor Cocktail were obtained from Roche (Mannheim, Germany), and lysozyme was purchased from Roth (Karlsruhe, Germany). Ni-IDA resin was obtained from Macherey-Nagel (Düren, Germany). All bacteria were obtained from Stratagene (Amsterdam, The Netherlands).

Purification of TcdA1 and TcdB2-TccC3

Escherichia coli BL21-CodonPlus cells were transformed with TcdA1-pET28a and grown in LB medium to an OD₆₀₀ of 0.6. Protein expression was induced by the addition of 25 μM isopropyl β-D-1-thiogalactopyranoside (IPTG) and the culture was further grown for 24 h at 22°C. For TcdB2-TccC3-pET28a, E. coli BL21 (DE3) cells were transformed and protein expression was induced with IPTG (25 μM). After 24 h at 28°C, the cells were harvested and used for further purification. Cells were resuspended in lysis buffer (300 mM NaCl, 20 mM Tris-HCl, pH 8.0, 1 mM dithiothreitol, 500 μM EDTA, and 10% glycerol) supplemented with DNase (5 μg/ml), lysozyme (1 mg/ml), and Protease Inhibitor Cocktail. After sonication, the cell lysate was incubated with Ni-IDA resin and loaded onto empty PD-10 columns. N-terminally His-tagged proteins were eluted with 500 mM NaCl, 20 mM Tris-HCl, pH 8.0, 0.05% Tween-20, 500 mM imidazole, and 5% glycerol. The protein-containing fractions of several columns were pooled and dialyzed against 100 mM NaCl, 50 mM Tris, pH 8.0, 0.05% Tween-20, and 5% glycerol.

Experiments with black lipid membranes

Black lipid bilayer membranes were formed from a 1% solution of diphytanoyl phosphatidylcholine (Avanti Polar Lipids, Alabaster, AL) in n-decane as described previously (21). The instrumentation consisted of a Teflon chamber with two aqueous compartments (cis and trans sides) connected by a small circular hole with a surface area of ∼0.4 mm². The membranes were obtained by painting the lipid solution across the hole. TcdA1 and TcdB2-TccC3 were always added from concentrated stock solutions to one side of the membrane in the black state (the cis side). The temperature was kept at 20°C throughout. The membrane current was measured with a pair of Ag/AgCl electrodes with salt bridges switched in series with a voltage source and in-house-made current amplifiers manufactured on the basis Burr Brown operational amplifiers. Zero-current membrane potential measurements were performed by establishing a salt gradient across membranes containing 100–1000 TcdA1 channels as described previously (22).

Channel size estimated from the effect of negatively charged groups at the channel

Net charges at the pore mouth result in substantial ionic strength-dependent surface potentials that attract cations and repel anions. Accordingly, either positive or negative charges influence both the single-channel conductance and zero-current membrane potential. Experimental evidence suggested that the TcdA1 channel contained negative net charges at the channel. A quantitative description of the effect of these point charges on the single-channel conductance may be obtained by the treatment proposed by Nelson and McQuarrie (23) and used by Trias and Benz (24) to describe the ionic-strength-dependent conductance of the cell-wall channel in Mycobacterium chelonae. It describes the effect of point negative (and positive) charge on the surface of a low dielectric membrane. The potential Φ created by a point charge q (in A·s), on the mouth of a channel with a radius r is given by

$$\Phi =\frac{2q·e^{-\frac{r}{l_D}}}{4\pi ·{\epsilon }_0·\epsilon ·r}$$ (1)

ε₀ (= 8.85 × 10⁻¹² F/m) and ε_r (= 80) are the absolute dielectric constant of vacuum and the relative constant of water, respectively, and l_D is the so-called Debye length, which controls the decay of the potential (and of the accumulated positively charged ions) in the aqueous phase:

$$l_D^2=\frac{\epsilon ·{\epsilon }_0·R·T}{2F^2·c}$$ (2)

where c is the bulk aqueous salt concentration, and R, T, and F (RT/F = 25.2 mV at 20°C) are the gas constant (R = 8.31 J/mol), the absolute temperature (T = 293K), and Faraday’s constant (F = 96500 A·s/mol), respectively.

The concentration of the monovalent cations near the point charge increases because of the negative potential. Its concentration, $c_0^{+}$, at the channel mouth is given by

$$c_0^{+}=c_0·e^{-\frac{\Phi ·F}{R·T}}$$ (3)

The cation concentration $c_0^{+}$ at the mouth of the pore can now be used for the calculation of the effective conductance-concentration curve:

$$G\left(c\right)=G_0·c_0^{+}$$ (4)

where G₀ is the concentration-independent conductance of the channel.

Binding experiments

The binding of TcdB2-TccC3 to the TcdA1 channel was investigated with titration experiments similar to those performed previously to study the binding of 4-aminoquinolones to the C2II and PA₆₃ channels, and edema factor and lethal factor to the PA₆₃ channel, in single- or multichannel experiments (11,26,27). The aqueous phase always contained 150 mM KCl buffered with 10 mM MES-KOH to a pH of 5.5–6. The TcdA1 channels were reconstituted into lipid bilayers. The reconstitution process of the TcdA1 channels proceeded until the rate of channel insertion in the membranes was very small (up to several hours). Then small amounts of a concentrated solution of TcdB2-TccC3 were added to the cis side of the membranes, with stirring to allow equilibration. The results of the titration experiments (i.e., blockage of the channels as a function of the concentration of the enzymatic components) were examined using the Langmuir adsorption isotherm (26,28):

$$\mathrm{\%}\text{of}\text{fraction}\text{of}\text{blocked}\text{channels}=\frac{\left(G_{\max }-G\left(c\right)\right)}{G_{\max }}=\frac{K·c}{\left(K·c+1\right)}$$ (5)

where K is the stability constant for binding of the enzymatic components TcdB2-TccC3 of the Tc toxins to the TcdA1-channels. The half-saturation constant K_s is given by the inverse stability constant 1/K. In case of only partial blockage of the TcdA1 channels induced by binding of TcdB2-TccC3 (i.e., the channels conduct still ions after binding, albeit at a much lower rate), Eq. 5 has to be modified to take the degree of blockage into account:

$$\mathrm{\%}\text{of}\text{degree}\text{of}\text{channel}\text{blocking}=\frac{\left(G_{\max }-G\left(c\right)\right)}{G_{\max }}=\text{A}·\frac{K·c}{\left(K·c+1\right)}$$ (6)

Where A represents the percentage of maximum blockage of the TcdA1 channel by TcdB2-TccC3 (∼70–80% in the binding experiments).

Results

Channel formation by TcdA1

In the first set of experimental conditions, we studied the interaction of TcdA1 with lipid bilayer membranes made of diphytanoyl phosphatidylcholine/n-decane. At low concentration (in the nanomolar range), TcdA1 formed ion-permeable channels in the membranes. However, channel formation was very slow under these conditions irrespective of TcdA1 concentration, which means that channel reconstitution could last for 3–4 h until approximate saturation was achieved, whereas for other channel-forming binding components of AB-like toxins, such as anthrax PA, C. botulinum C2II, or C. perfringens iota b, saturation of channel formation was obtained in reconstitution experiments within ∼30 min to 1 h (8,9,27). Fig. 1 A shows a single-channel recording of current fluctuations of a diphytanoyl phosphatidylcholine/n-decane membrane observed with 6 nM TcdA1 in 150 mM KCl solution. The single-channel conductance was on average ∼125 pS at a membrane potential of 20 mV and the channels had long lifetimes of at least 5 min under the conditions of Fig. 1 A. Fig. 1 B demonstrates that the fluctuations were fairly homogeneous and centered around 125 pS (>80% of all single events). The formation of channels by TcdA1 in lipid bilayer membranes was not a rare event. It is noteworthy that with 10 nM TcdA1, >1000 channels were formed in a diphytanoyl phosphatidylcholine/n-decane membrane with a surface area of ∼0.4 mm² within a time course of several hours.

Figure 1.

Single-channel recording and distribution of TcdA1 from P. luminescens. (A) Approximately 10 min after the formation of the membrane from diphytanoyl phosphatidylcholine/n-decane, 6 nM TcdA1 was added to the aqueous phase on one side of the membrane (the cis side). The aqueous phase contained 150 mM KCl (10 mM MES-KOH, pH 6). The applied membrane potential was 20 mV (T = 20°C). Note that the current noise increased after the insertion of the seventh channel, presumably caused by the reconstitution of a fuzzy channel. (B) Histogram of the probability of the occurrence of certain conductivity units observed with membranes formed of diphytanoyl phosphatidylcholine/n-decane in the presence of TcdA1 from P. luminescens. The aqueous phase contained 150 mM KCl, 10 mM MES-KOH, and 6 nM TcdA1 from P. luminescens. The applied membrane potential was 20 mV (T = 20°C). The average single-channel conductance was 125 pS for 95 single-channel events. The data were collected from five different membranes.

Single-channel analysis

The lipid bilayer system allows excellent control of the conditions on both sides of the membranes. Thus, it is possible to study the properties of the TcdA1 channel as a function of a variety of parameters, such as salt concentration and ionic composition of the salts. The results of these experiments are summarized in Table 1. At high KCl concentration, the channels formed by TcdA1 showed a linear dependence on the KCl concentration in the bulk aqueous phase (see Table 1). At low ionic strength, we observed a considerable variation from linearity, by a factor of ∼10 at a KCl concentration of ∼10 mM. The slope of the conductance versus concentration curves on a double logarithmic scale was ∼0.5, which indicated the influence of point net charges in or near the TcdA1 channels on ion flow through the channel. This suggested that, in similarity to channels formed by anthrax PA and C2II, the channel contains a cluster of negatively charged groups. We previously showed that the effect of these groups on the channel conductance allows one to obtain a rough estimate of the channel size of TcdA1 (9,24,27) (see Discussion).

Table 1.

Average single-channel conductance, G, of the channel formed by TcdA1 of P. luminescens in different salt solutions, and comparison with channel formation by C2II of the C2 toxin of Clostridium botulinum

Salt	Concentration, c (M)	Single-channel conductance, G (pS)
		TcdA1	C2II
KCl	0.01	24 ± 5	12
	0.03	47 ± 5	25
	0.1	95 ± 10	55
	0.15	125 ± 9	n.m.
	0.3	190 ± 23	80
	1.0	495 ± 40	150
	3.0	1420 ± 150	380
LiCl	1.0	265 ± 25	60
K+-acetate	1.0	360 ± 21	120

The membranes were formed of diphytanoyl phosphatidylcholine dissolved in n-decane. The aqueous solutions were buffered with 10 mM MES-KOH and had a pH of 6. The applied voltage was 20 mV and the temperature was 20°C. The average single-channel conductance (± SD) was calculated from at least 100 single events from at least three individual membranes. The single-channel conductance of the C2II binding component of the Clostridium botulinum C2 toxin is given for comparison (9). c, concentration of the aqueous salt solution; n.m., not measured.

Single-channel experiments were also performed with other salt solutions. These experiments were performed to gain some insight into the biophysical properties of the TcdA1 channel and its selectivity. The results are also included in Table 1 and show that cations had a more substantial influence on the single-channel conductance than anions. This result is consistent with the assumption that the TcdA1 channel is preferentially cation selective. The ionic selectivity of the cations was K⁺ > Li⁺, which means that the permeability of the cations through the channels was consistent with their mobility sequence in the aqueous phase. This suggested that the TcdA1 channel is water-filled, and inside has only small field strength and no narrow-spaced selectivity filter (i.e., no binding site). In agreement with this, we observed that the single-channel conductance also decreased in potassium acetate somewhat as compared with KCl, indicating that anions also can enter the channel and influence its conductance.

In additional experiments, we investigated the influence of pH on conductance of the TcdA1 channel. Interestingly, we found the same slow reconstitution rate for the variation of pH between 5.0 and 7.0. However, the channel conductance was highly dependent on the aqueous pH and increased from pH 5.0 to 7.0 by a factor of ∼2 from 78 pS to 165 pS in 150 mM KCl, indicating that charges inside the channel may change their sign (see Table 2).

Table 2.

Average single-channel conductance, G, of channels formed by TcdA1 in 150 mM KCl solutions at different pH values

Electrolyte	pH	Single-channel conductance (pS)
KCl	5.0	78 ± 9
KCl	6.0	125 ± 9
KCl	7.0	165 ± 15

The membranes were formed of diphytanoyl phosphatidylcholine in n-decane. The aqueous solutions were buffered with 10 mM MES-KOH or 10 mM Tris-HCl and had the given pH. The applied voltage was 20 mV and the temperature was 20°C. The average single-channel conductance (± SD) was calculated from at least 100 single events from at least three individual membranes.

Selectivity of the channel formed by TcdA1

The selectivity of the TcdA1 channel was measured in zero-current membrane potential measurements in the presence of salt gradients. After a large number of channels were incorporated in membranes bathed in 100 mM KCl, a fivefold salt gradient was established across the membranes by addition of small amounts of concentrated KCl solution to one side of the membrane. In all cases, the more diluted side of the membrane became positive, indicating preferential movement of cations through the TcdA1 channel (see Table 3), i.e., it is cation selective as suggested from the single-channel data (see Table 1). Analysis of the zero-current membrane potentials using the Goldman-Hodgkin-Katz equation (22) showed that the permeability ratio P_cation divided by P_anion followed the aqueous mobility of the ions (see Table 3), which is consistent with the assumption that TcdA1 forms a water-filled channel with some preference for cations. It is noteworthy that in contrast to TcdA1, the binding components C2II and anthrax PA form highly cation-selective channels, with a permeability ratio P_cation and P_anion of >10 for KCl (9,27).

Table 3.

Zero-current membrane potentials, V_m, of diphytanoyl phosphatidylcholine/n-decane membranes in the presence of TcdA1 measured for a fivefold gradient of different salts

Salt	Vm/mV	Pcation/Panion	n
KCl	20.1 ± 1.7	3.6 ± 0.5	4
LiCl	11.4 ± 0.2	2.0 ± 0.3	3
KCH3COO (pH 7)	27.5 ± 0.7	6.6 ± 0.5	3

V_m (± SD) is defined as the difference between the potential at the diluted side (100 mM) and the potential at the concentrated side (500 mM). The pH of the aqueous salt solutions was 6 unless otherwise indicated; T = 20°C. The permeability ratio P_cation/P_anion (± SD) was calculated using the Goldman-Hodgkin-Katz equation (22). n is the number of individual experiments used for the calculation of the zero-current membrane potentials and the permeability ratio.

The TcdA1 channel is voltage dependent in an asymmetric manner

In single-channel recordings, the TcdA1 channel exhibited some flickering at higher voltages, i.e., it showed rapid transitions between open and closed configurations. This could be caused by its voltage-dependent closing, and therefore we increased the voltage across the membranes in single- and multichannel recordings. Fig. 2 A shows an experiment of the latter type. TcdA1 was added in a concentration of 10 nM to one side of a black diphytanoyl phosphatidylcholine/n-decane membrane (the cis side) and the conductance increase was followed for ∼2 h. After this time, the reconstitution rate of TcdA1 decreased significantly, and we applied different negative and positive potentials (with respect to the cis side) to the membrane starting from ±20 mV (left side of Fig. 2 A). We then repeated the experiment with ±40, ±60, ±80, and ±100 mV. Fig. 2 A shows the results obtained with positive potentials at the trans side (upper traces) followed by negative potentials at the trans side of the membrane (lower traces). Only for positive potentials at the trans side did the membrane current decrease in an exponential fashion starting with ∼+40 mV (see Fig. 2 B). For negative potentials at the trans side (corresponding to positive potentials at the cis side, mimicking the in vivo situation), the current did not decrease even when the membrane potential was as high as 150 mV (data not shown). This result indicated full asymmetric insertion of TcdA1 into the membranes when it was added to only one side. The addition of the protein to both sides of the membrane resulted in a symmetric response to the applied voltage (data not shown).

Figure 2.

Asymmetric current response of the TcdA1 channel upon application of different negative and positive potentials to the trans side of a membrane. (A) The membrane was made of diphytanoyl phosphatidylcholine/n-decane bathed in 0.1 M KCl, 10 mM MES-KOH, pH 6. The cis side contained 10 nM TcdA1 (T = 20°C). Note that the membrane current decreased only when the cis side (i.e., the side of the addition of TcdA1) was negative. (B) Ratio of the conductance G at a given membrane potential (V_m) divided by the conductance G₀ at 20 mV as a function of the membrane potential V_m. The closed points show measurements obtained after 10 nM TcdA1 was added to the cis side of membranes. The voltage refers to that observed on the cis side of the membrane. The aqueous phase contained 0.1 M KCl, 10 mM MES-KOH, pH 6. The membranes were formed from diphytanoyl phosphatidylcholine/n-decane. T = 20°C. Means of four experiments are shown.

The data obtained from the experiment in Fig. 2 A and similar experiments were analyzed in the following way: the membrane conductance (G) as a function of voltage, V_m, was measured when the opening and closing of channels reached an equilibrium, i.e., after the exponential decay of the membrane current following the voltage step V_m. G was divided by the initial value of the conductance (G₀, which was a linear function of the voltage) obtained immediately after the onset of the voltage. The data of Fig. 2 B correspond to the asymmetric voltage dependence of TcdA1 (mean of three membranes) when the protein was added to the cis side. The results indicate full orientation of the TcdA1 channel when the protein was only added to one side of the membrane (the cis side). Furthermore, it is clear from the data of Fig. 2 A that the TcdA1 channel shows a slightly asymmetric current voltage curve, because the currents obtained for negative potentials at the cis side are somewhat higher than those for positive potentials.

Blockage of TcdA1 channels by TcdB2-TccC3

Previous studies have presented evidence that the enzymatic components of certain AB-type toxins, such as anthrax and C2 toxin, enter target cells via permeation through the channels formed by the binding components (10,25). To test whether the enzymatic component TcdB2-TccC3 also interacts with the channels formed by TcdA1, we performed multichannel experiments in the presence of TcdA1. The channels formed by the binding component were reconstituted into membranes until the reconstitution rate slowed down to zero (which sometimes took several hours) and the membrane contained many TcdA1 channels. Fig. 3 shows an experiment of this type. TcdA1 was added in a concentration of 10 nM to one side of a black membrane bathed in 150 mM KCl, 10 mM MES-KCl pH 6. After ∼3 h, the membrane conductance was approximately stationary. At this time, 0.4 nM TcdB2-TccC3 was added to the same side of the membrane (the cis side, arrow in Fig. 3). This led to a substantial decrease of the membrane conductance, indicating a blockage of the TcdA1 channels by TcdB2-TccC3 (see Fig. 3). Further additions of TcdB2-TccC3 resulted in a further decrease of the TcdA1-mediated membrane conductance. The data shown in Fig. 3 and obtained in similar experiments were analyzed using the Langmuir adsorption isotherm (Eqs. 5 and 6) (26,28), as shown in Fig. 4. The fit curve (solid line in Fig. 4, using Eq. 6). and a maximum blockage of ∼70% of the individual channel) corresponds to a stability constant K of (1.76 ± 0.25) × 10⁹ M⁻¹ for TcdB2-TccC3 binding to TcdA1 (half-saturation constant K_S = 0.57 nM). The stability constant K of the binding of TcdB2-TccC3 to TcdA1 channels was averaged out of at least three individual experiments, resulting in K = (1.8 ± 0.4) × 10⁹ M⁻¹ (half-saturation constant K_s = 0.56 nM). The maximum blockage of the TcdA1 channels was ∼75% (i.e., they conducted ∼25% of the current without binding of TcdB2-TccC3; see also below).

Figure 3.

Titration of TcdA1-induced membrane conductance with TcdB2-TccC3. The membrane was formed from diphytanoyl phosphatidylcholine/n-decane. The aqueous phase contained 150 mM KCl, 10 mM MES-KCl pH 6, and 10 nM TcdA1 at the cis side. TcdB2-TccC3 was added to the cis side at the concentrations shown at the top of the figure while stirring. The temperature was 20°C and the applied voltage was 20 mV at the cis side. For further explanations, see text.

Figure 4.

Langmuir adsorption isotherm of the inhibition of TcdA1-induced membrane conductance by TcdB2-TccC3. The data of Fig. 3 were analyzed using Eq. 6. The solid curve corresponds to a stability constant K for TcdB2-TccC3 binding to TcdA1 channels of (1.76 ± 0.25) × 10⁹ M⁻¹ (K_s = 0.57 nM) and a maximum blockage of the TcdA1 channels by A = ∼70%. For further explanations, see text.

The data of Figs. 3 and 4 suggested that blockage of TcdA1 channels by TcdB2-TccC3 was <100%. Analysis of the titration experiments with Eq. 6 demonstrated a maximum blockage of the single channel by ∼75%. To check whether this could also be detected in single-channel experiments, we performed reconstitution experiments with diphytanoyl phosphatidylcholine/n-decane membranes in 1 M KCl, 10 mM MES-KCl pH 6 in the presence of 1 nM TcdA1. After reconstitution of channels, when the reconstitution rate was very low, the stirrer was switched on in the membrane cell, which caused a considerable current noise. TcdB2-TccC3 was added in an extremely small concentration (0.1 nM) to the cis side of the membrane (Fig. 5, arrow). Subsequently, the channels closed in a defined fashion. Analysis of the closing events revealed a conductance of the events of 390 ± 15 pS at a voltage of 20 mV, which is ∼79% of that of the open state (495 pS; see Table 1). This means that binding of TcdB2-TccC3 to the TcdA1 channel led to only a partial inhibition of the current, as suggested by the titration experiments.

Figure 5.

Blockage of the TcdA1-induced conductance with TcdB2-TccC3 on the single-channel level. Approximately 20 TcdA1-channels were reconstituted in a black diphytanoyl phosphatidylcholine/n-decane membrane from the cis side before the start of the experiment. The membrane cell was then stirred, which caused a considerable current noise. At the position indicated by the arrow, 0.1 nM TcdB2-TccC3 was added to the cis side. Subsequently, the current decreased in a stepwise fashion caused by blockage of the TcdA1 channels with TcdB2-TccC3. The closing events had on average a conductance of 390 ± 15 pS. The aqueous phase contained 1 M KCl, 10 mM MES-KCl pH 6; the applied voltage was 20 mV at the cis side.

Discussion

TcdA1 of P. luminescens forms ion-permeable channels in lipid bilayer membranes

TcdA1 of P. luminescens and related proteins of other gram-negative and gram-positive bacteria are components of a novel class of toxins, the Tc toxin complexes, which act preferentially against insects (29–32). The toxins are composed of different components (TcdA1, TcdB2, and TccC3 or TccC5) (13,33,34). The biologically active components TccC3 and TccC5 are ADP-ribosyltransferases that selectively modify unusual amino acids within actin and Rho proteins, respectively (13). Whereas arginine residues are modified by numerous ADP-ribosyltransferases, such as cholera toxin (35) and the family of binary actin-ADP-ribosylating toxins (20), cysteine and asparagine residues are modified by pertussis toxin (36,37) and C3-like toxins (38), respectively. TccC3 ADP-ribosylates threonine-148 in actin, resulting in its polymerization, whereas TccC5 ADP-ribosylates Rho proteins at glutamine-61/63, inducing their activation (13). The concerted action of both toxins induces extensive intracellular polymerization and clustering of actin. TcdA1 and TcdB2 presumably are involved in importing enzymatically active components into the target cells. The toxin complexes have high molecular masses (>1.4 MDa) (14,29). It was previously demonstrated that the component TcdA1 plays a crucial role in the uptake of enzymatic components into target cells (13). More recently, the structure of the TcdA1 oligomer was determined by cryo-EM and single-particle analysis at 6.3 Å resolution (14). In this study, we demonstrate that TcdA1 forms well-defined ion-permeable channels in lipid bilayer membranes, which could represent transport channels for the effectors. In similarity to channels formed by binding components of AB-type toxins such as anthrax PA (7,27,39) and C2II (9), TcdA1 appeared to be cation selective, as indicated by single-channel and selectivity measurements. Channel formation always appeared to be slow in lipid bilayer membranes independently of the TcdA1 concentration, which means that it took a long time for saturation of the reconstitution process to be reached. This could be caused by the relatively high stability of TdcA1 in the aqueous phase and the time needed to form the pentamer from the five monomers. However, it may also be due to the high molecular mass of the pentamer (∼1.4 MDa), which means that many lipid molecules have to be pushed aside to provide the space for this huge aggregate.

Negative point charges at the channel mouth influence the flux of charged molecules through the TcdA1 channel

The cation selectivity of TcdA1 is obviously based on the presence of point charges at the channel mouth, because the single-channel conductance was not a linear function of the bulk conductivity of the KCl solutions, as the data of Table 1 clearly show. Thus, we tried to fit the single-channel conductance of TcdA1 as a function of the bulk aqueous KCl concentration using the Nelson and McQuarrie (23) formalism as we did previously for channels formed by other binding proteins (9,27) (see also Eqs. 1–4). A best fit of the concentration dependence of the single-channel conductance shown in Fig. 6 was obtained assuming that 1.6 negatively charged groups (q = −2.56 × 10⁻¹⁹ A·s) are located at the pore mouth and that its radius is ∼1.1 nm. This is very close to the value reported for the cryo-EM model obtained at 6.3 Å resolution with a radius of 1.5 nm at the entrance and 0.75 nm at the bottleneck of the channel, which agrees nicely with single-channel measurements obtained using nonelectrolytes of different sizes (14). The conductance of the channel in the absence of the negatively charged groups is 500 pS/M. The results of Fig. 6 show that a reasonable fit of the experimental data is achieved for these parameters. Fig. 6 also demonstrates that the influence of the surface charges is rather small at high ionic strength. When the single-channel conductance is corrected for the presence of point charges at the channel mouth according to Eq. 4, it is a linear function of the aqueous salt concentration (broken line in Fig. 6). It is noteworthy that a similar treatment of the C2II channel, taken from Schmid et al. (9), showed that C2II formed a much smaller channel (see Fig. 6). The smaller channel size of C2II resulted in a much stronger effect of the negative charges on channel conductance at low ionic strength (Fig. 6). The parameters of the fit of the single-channel conductance of C2II as a function of the KCl concentration in terms of the Nelson and McQuarrie (23) formalism were q = −1.4 elementary charges (q = 2.22 × 10⁻¹⁹ A·s), r = 0.75 nm and G₀ = 150 pS/M, indicating that the C2II channel indeed has a much lower conductance than the TcdA1 channel.

Figure 6.

Single-channel conductance of TcdA1 as a function of the KCl concentration. The membranes were formed from diphytanoyl phosphatidylcholine/n-decane. The aqueous phase was buffered with 10 mM MES-KOH, pH 6, and contained 6 nM TcdA1 added to the cis side of the membrane. The temperature was 20°C and the applied voltage was 20 mV at the cis side. The fit of the experimental data was achieved using Eqs. 1–4 assuming the parameters q = 2.56 × 10⁻¹⁹ A·s and r = 1.1 nm. The data for C2II were taken from Schmid et al. (9). Note that the conductance of TcdA1 was similar to that of C2II, and not a linear function of the aqueous salt solution. For further explanations, see text.

The negative potential at the mouth of the channel has important implications for the function of the TcdA1 channel of P. luminescens. At a concentration of 150 mM KCl or NaCl, the potential is ∼−14 mV at the channel mouth (as calculated from the Nelson and McQuarrie formalism (23)). This means that the concentration of monovalent cations is increased there to 263 mM (bulk concentration 150 mM, calculated according to c₀⁺ = c exp (−ϕ·F/(R·T))), whereas the concentration of monovalent anions is decreased to 85 mM (bulk concentration 150 mM, calculated according to c₀− = c exp (ϕ·F/(R·T)). This means that under physiological conditions, the channel conducts cations ∼3-fold better than anions of the same aqueous mobility without being really selective for cations due to the presence of a binding site for cations. Similar considerations apply to the discussion of the zero-current membrane potentials, i.e., only part of the full bulk aqueous gradient drops across the channel and influences its selectivity. It is noteworthy that other binding protein channels, such as anthrax PA and C2II, are also influenced by point charges (7,9,27).

The TcdA1 channel is voltage gated

The TcdA1 channel showed voltage-dependent gating. Starting with approximately −20 to −30 mV applied to the cis side of the membrane (the same side to which the 285 kDa protein was added), the current through the channels decreased in an exponential time-dependent manner (see Fig. 2 A). For opposite potential at the cis side of the membrane, the current was not influenced up to voltages of 100 mV and even higher voltages (data not shown). This result indicated asymmetric insertion of the TcdA1 into the membrane. Possibly a large hydrophilic part of the cell wall channel is exposed to the aqueous phase on the cis side of the membrane (i.e., the side of the addition of the protein), whereas the more-hydrophobic channel-forming domain crosses the membrane and leads to a transmembrane channel. This asymmetric distribution of the TcdA1 channel would result in an asymmetric response toward the sign of the membrane potential. This result could mimic the in vivo situation in which the bulky hydrophilic part of the TcdA1 channel is exposed to the surface of the cytoplasmic membrane. Then the TcdA1 channel should be open for negative potentials at the inside of the cytoplasmic membrane and at the surface of the endosomes, representing the in vivo situation. For potentials of opposite sign, the channels would close. Interestingly, the channels formed by binding proteins of anthrax and C2 toxin also show voltage-dependent gating for negative potentials at the cis side (7,9).

Conclusions

Tc toxins contain channel-forming components

In this study we have demonstrated that a component of the Tc toxin system of P. luminescens forms channels in lipid bilayer membranes. Channel formation occurred frequently at very low TcdA1 concentrations, but very slowly, which means that reconstitution events occurred over several hours and it was difficult to reach equilibrium. The reason for this is not clear because the TcdA1 pentamers seem to be preformed in the aqueous phase. On the other hand, the pentamers have an extremely high molecular mass of ∼1.41 MDa. A considerable energy may be needed to insert such a big complex into the membranes, where many lipid molecules have to be replaced to allow the insertion of the TcdA1 pentamer into the membrane.

Channel formation seems to be a prerequisite for toxin transport into target cells. This has been shown for a variety of AB-type toxins, such as anthrax, C2, and iota toxins (7–10). However, channel formation has also been demonstrated for diphtheria toxin (1–3) and different clostridial neurotoxins (4–6). All of these toxins form small channels with a conductance of <100 pS in 150 mM KCl, which means that they definitely have a much smaller conductance than the TcdA1 channel. Nevertheless, in all of these cases, channel formation is essential for toxin delivery, which is also the case for Tc toxins. It seems that the Tc toxin system provides a completely different delivery system compared with the other AB-type toxins, because the binding components of anthrax, C2, and iota toxins form all β-hairpin, and the β-barrel cylinder formed by these binding components contains seven or eight monomers, comprising a total of 14 or 16 β-strands (7–10). In fact, a recent preliminary structural analysis by cryo-EM suggested that the TcdA1 pentamer is composed entirely of α-helical structures (14). How this novel type of transport system functions to translocate the enzyme component into the cytosol of target cells remains to be clarified. Moreover, further studies are needed to elucidate a possible role of the TcB component in translocation.