The equinatoxin N-terminus is transferred across planar lipid membranes and helps to stabilize the transmembrane pore Katarina Kristan 1 , Gabriella Viero 2 , Peter Mac ˘ ek 1 , Mauro Dalla Serra 2 and Gregor Anderluh 1 1 Department of Biology, Biotechnical Faculty, University of Ljubljana, Ljubljana, Slovenia 2 ITC-CNR Institute of Biophysics, Unit at Trento, Trento, Italy Pore-forming toxins (PFTs) comprise one of the most widespread groups of natural toxins. They have the unusual characteristic of existing in two different states: they are synthesized as soluble monomers, which spontaneously insert into cellular and model membranes to form transmembrane pores that are per- meable to various compounds [1,2]. The formation of transmembrane pores disrupts ion gradients, which leads to osmotic swelling and ultimately to cell death. PFTs differ in sequence and structure, but the major steps of pore formation are similar. The first step is attachment to the membrane surface, which is a non- specific or a specific process, with the receptor being a membrane protein or a lipid molecule; for example, cholesterol-dependent cytolysins from Gram-positive bacteria only bind to membranes including cholesterol [3]. Binding to the membrane is essential for oligomeri- zation, as it enables a high concentration and correct orientation of molecules on the host cell membrane. After initial attachment, PFTs undergo a series of Keywords actinoporin; equinatoxin; planar lipid membrane; pore-forming toxin; pore structure Correspondence M. Dalla Serra, ITC-CNR Institute of Biophysics, Unit at Trento, Via Sommarive 18, 38050 Povo, Trento, Italy Fax: +39 0461 810628 Tel: +39 0461 314156 E-mail: mdalla@itc.it G. Anderluh, Department of Biology, Biotechnical Faculty, University of Ljubljana, Vec ˘ na pot 111, 1000 Ljubljana, Slovenia Fax: +386 1 257 3390 Tel: +386 1 42 333 88 E-mail: gregor.anderluh@bf.uni-lj.si (Received 27 August 2006, revised 11 November 2006, accepted 21 November 2006) doi:10.1111/j.1742-4658.2006.05608.x Equinatoxin II is a cytolytic protein isolated from the sea anemone Acti- nia equina. It is a member of the actinoporins, a family of eukaryotic pore- forming toxins with a unique mechanism of pore formation. Equinatoxin II is a 20 kDa cysteineless protein, with sphingomyelin-dependent activity. Recent studies showed that the N-terminal region of the molecule requires conformational flexibility during pore formation. An understanding of the N-terminal position in the final pore and its role in membrane insertion and pore stability is essential to define the precise molecular mechanism of pore formation. The formation of pores and their electrophysiologic char- acteristics were studied with planar lipid membranes. We show that amino acids at positions 1 and 3 of equinatoxin II are exposed to the lumen of the pore. Moreover, sulfhydryl reagents and a hexa-histidine tag attached to the N-terminus revealed that the N-terminus of the toxin extends through the pore to the other (trans) side of the membrane and that negat- ively charged residues inside the pore are crucial to define the electrophysio- logic characteristics of the channel. Finally, we detected a new, less stable, state with a lower conductance by using a deletion mutant in which the first five N-terminal amino acids were removed. We propose that the first five amino acids help to anchor the amphipathic helix on the trans side of the membrane and consequently stabilize the final transmembrane pore. Abbreviations EqtII, equinatoxin II; His 6 –EqtII, fusion protein with hexa-histidine tag attached to the N-terminus of equinatoxin II; MTS, methane- thiosulfonate; MTSEA + , (2-aminoethyl)-methanethiosulfonate hydrobromide; MTSES – , sodium (2-sulfonato-ethyl)-methanethiosulfonate; MTSET + , (2-trimethylamonium)-ethyl methanethiosulfonate; PFT, pore-forming toxin; PLM, planar lipid membrane; SM, sphingomyelin. FEBS Journal 274 (2007) 539–550 ª 2006 The Authors Journal compilation ª 2006 FEBS 539 conformational changes that expose previously hidden hydrophobic parts for the interaction with the membrane lipids. PFTs are commonly divided into two subgroups according to structural elements of the transmembrane pore. The pores of b-PFTs are transmembrane b-barrels formed by interconnected b-hairpins. Examples include the well-known Staphylo- coccus aureus a-toxin, cholesterol-dependent cytolysins and anthrax toxin protective antigen [4,5]. On the other hand, a-PFTs build pores from amphipathic a-helices [6]. The most notable examples are colicins [7] and actinoporins [8]. Transmembrane b-barrel pores are structurally stable, whereas pores formed by a-helices are unstable, and consequently there is less structural information available. Because of the above- mentioned properties, PFTs comprise a useful and unique model with which to study protein unfolding on the surface of the lipid bilayer [9], oligomerization in a hydrophobic environment and membrane protein structure [10], protein–protein interactions on the surface of the membrane [11], etc. In addition, PFTs are particularly interesting as potential tools in biotechnology; that is, they can be used for selective killing of cancer cells, with built-in biological ‘triggers’ that activate in response to specific biological stimuli [12,13], or they can be used in biosensor technology [14]. Actinoporins, cytolytic toxins synthesized by sea anemones, comprise a unique group of PFTs [8]. They comprise a group of 20 kDa, cysteineless proteins, whose activity depends on the presence of sphingomye- lin (SM) in the membrane. The most studied represen- tative actinoporins are equinatoxin II (EqtII), isolated from the sea anemone Actinia equina, and sticho- lysin II, from Stichodactyla helianthus. Actinoporins form cation-selective pores approximately 2 nm in diameter on the surface of the target cell [15,16], lead- ing to cell lysis through colloid osmotic shock. The three-dimensional structures of EqtII and sticholysin II monomers were determined in solution [17–19]. The molecule is composed of a tightly folded hydrophobic b-sandwich core, flanked on two faces by a-helices (Fig. 1A). The first 30 N-terminal amino acids, inclu- ding an amphipathic a-helix, form the only part that can detach from the core of the molecule without disrupting the general fold of the protein [17,19]. Recently, it has been shown that this is the only part of the molecule that undergoes major conformational changes during pore formation, and that its flexibility is essential for formation of the final pore [20,21]. Act- inoporin pore formation proceeds in distinct steps. The initial attachment to the membrane is achieved by a cluster of exposed aromatic amino acids situated on the broad loops at the bottom of the molecule and the C-terminal a-helix [20,22,23], and by a recently defined phosphorylcholine-binding site [19]. In the next step, the N-terminal a-helix detaches from the core of the molecule and inserts into the lipid–water interface, where it lies flat on the membrane [19,20,24]. Finally, four toxin monomers oligomerize and form the pore by inserting the N-terminal a-helix through the mem- brane [24]. The final functional pore is thus composed of four amphipathic helices from four monomers [15,16,25] and most probably also by membrane lipids in a so-called toroidal pore arrangement [26,27]. The pores formed by actinoporins have not yet been directly visualized. Most previous experiments were focused on the N-terminal a-helix, which extends from Ser15 to Leu26 in solution [17,18] and from Asp10 to Asn28 in a hydrophobic membrane environment [24]. The question remains of how the region that comprises residues 1–10 is organized in the final pore, and what the nature of the contacts is between the monomers, which should stabilize the toroidal pore [19]. Thus, the purpose of this work was to gain further insight into the structure of the EqtII pore, especially the topology of the first five N-terminal amino acids. We demon- strated that the N-terminus is positioned on the trans side of the membrane. Furthermore, the amino acids at positions 1 and 3 are exposed to the ion conductive pathway, and the N-terminus helps to stabilize the final pore. Results Characterization of transmembrane channels formed by EqtII, single cysteine and deletion mutants In this study, we used planar lipid membranes (PLMs) and three N-terminal EqtII mutants to study the topo- logy of the N-terminus in the final transmembrane pore (Fig. 1C). We chose to study the S1C mutant, as the modification of a thiol group with methanethiosulf- onate (MTS) reagents allows incorporation of a posit- ive charge by using (2-aminoethyl)-MTS hydrobromide (MTSEA + ) and (2-trimethylamonium)-ethyl MTS (MTSET + ), or a negative charge by using sodium (2-sulfonato-ethyl)-MTS (MTSES – ). In order to clarify the effect of the first five residues on the electrophysio- logic properties and stability of the pore, the deletion mutant D5 was used. This mutant lacks one negative charge (Asp3) and three hydrophobic residues (Fig. 1C,D). Finally, a His 6 –EqtII variant contains a hexa-histidine tag and prolongs the N-terminus of EqtII for 13 amino acid residues, adding a strong positively charged region to the N-terminal helix at Topology of equinatoxin II N-terminus K. Kristan et al. 540 FEBS Journal 274 (2007) 539–550 ª 2006 The Authors Journal compilation ª 2006 FEBS pH 5.5; at this pH, the majority of the histidines should be protonated [28] (Fig. 1C). All of the mutants were produced in Escherichia coli and purified to homogeneity as shown by SDS ⁄ PAGE gels (Fig. 1B). The hemolytic activity of S1C and D5 was as observed previously [24,29]. The addition of the His-tag to the N-terminus of EqtII decreased the hemolytic activity (c wt 50 ¼ 0.21 lgÆmL )1 ± 0.03, c HisÀEqtII 50 ¼ 0.98 lgÆmL )1 ± 0.04; n ¼ 3–5 ± SD) (c 50 is the concentra- tion of a protein that produces 50% of the maximal rate of hemolysis). Deletion mutant D5, S1C, chemically modified ver- sions of S1C (S1C-MTSEA + , S1C-MTSET + , S1C- MTSES – , in which S1C was chemically modified with MTSEA + , MTSET + and MTSES – , respectively), and His 6 –EqtII were able to form pores in PLMs at final concentrations of 1–5 nm (Figs 2 and 4). EqtII and its mutants formed pores in PLMs with a broad conductance distribution: 308 pS (wild-type), 329 pS (S1C), 349 pS (S1C-MTSEA + ), 358 pS (S1C- MTSET + ), 247 pS (S1C-MTSES – ) and 256 pS (D5) (Fig. 3A). D5 frequently showed two different types of pore. One was similar to the wild-type, and the other had a lower conductance and was less stable (Fig. 2A, bottom trace, arrows). This behavior was not seen in the wild-type or other mutants used in this study, sug- gesting that it was a peculiarity of the deletion mutant. These lower-conductance pores typically had conduct- ances of 100–150 pS (Fig. 2B) and remained open for a few seconds to 60 s (Fig. 2A). After 5–10 min, when pores with higher conductance opened, such events were no longer observed, probably due to the typical total noise of EqtII multichannel recordings. All mutants showed cation selectivity, but to different extents (Fig. 3B). Mutation of Ser1 to Cys did not affect the selectivity significantly (P + ⁄ P À WT ¼ 9.08; P + ⁄ P À S1C ¼ 10.14; p ¼ 0.321) (P + and P ) are the permeability of cation and anion, here K + and Cl ) respectively). The addition of a negative charge (S1C- MTSES – ) leads to a significant increase in cationic selec- tivity (P + ⁄ P À MTSESÀ ¼ 15.1; p ¼ 0.021). On the other hand, the addition of positive charge (S1C-MTSEA + , S1C-MTSET + ) or the deletion of a negative charge in D5 caused a shift to less cationic selectivity AD B C Fig. 1. Structure of EqtII and alignments of actinoporin N-terminal sequences, EqtII and mutants used in the study. (A) Three-dimen- sional structure of EqtII (Protein Data Bank code 1IAZ) with its N-terminal amphipathic a-helix shown in black. (B) SDS ⁄ PAGE gel of proteins used in this study. Approxi- mately 1 lg of each protein was resolved on 12% SDS ⁄ PAGE gel and stained with Coomassie Blue. (C) Alignment of the wild- type EqtII, mutants and fusion protein used. Negatively charged amino acids are in black. The position of the N-terminal amphipathic a-helix (amino acids 15–26) is shown above the alignment by letters h [17]. The region that was shown to be in an a-helical arrangement in the final pore is underlined [24]. An arrow denotes the thrombin clea- vage site. (D) The alignment of known N-ter- minal sequences of actinoporins compiled from the literature [8,45–49]. Hydrophobic amino acids (AGVLIFWYP) are shown in black, polar amino acids (TSMCNQ) are shown in dark gray, and charged amino acids (DEKHR) are shown in light gray. The numbering is according to EqtII. K. Kristan et al. Topology of equinatoxin II N-terminus FEBS Journal 274 (2007) 539–550 ª 2006 The Authors Journal compilation ª 2006 FEBS 541 (P + ⁄ P À MTSEAþ ¼ 5.07, p ¼ 0.0003; P + ⁄ P À MTSETþ ¼ 4.03, p ¼ 0.0002; P + ⁄ P À d5 ¼ 2.58, p < 0.0001) (Fig. 3B). To further evaluate the charge distribution through the pore and possibly the position of the N-terminus, the current–voltage characteristic (I ⁄ V) was studied (Fig. 3C). The symmetry of the I ⁄ V curve for S1C, i.e. I + ⁄ I – % 1 (Fig. 3C, inset), suggests that the distribu- tion of the charges along the lumen of the mutated pore was not significantly different from that of the wild-type pore. The addition of a negative charge on S1 (S1C-MTSES – ) slightly decreased the I + ⁄ I – ratio (Fig. 3C, inset). On the other hand, the addition of a positive charge (S1C-MTSET + ) strongly increased the slope of the I + ⁄ I – ratio vs. applied voltage curve (Fig. 3C, inset), which is consistent with the location of that charge on the trans side. A similar strong asymmetry in the I ⁄ V curve was also observed for D5, which lacks the negative charge at position 3. The position of the N-terminus in the final pore His 6 –EqtII was used to analyze in a more direct way the position of the N-terminus in the final pore, i.e. to which side of the membrane the N-terminus extends. His 6 –EqtII pores are slightly less conductive than the wild-type [G ¼ 204 ± 36.6 pS (p ¼ 0.004; n ¼ 21 average ± SD) (G ¼ conductance), compared to 308 pS for the wild-type]. The selectivity was not chan- ged (P + ⁄ P À His6ÀEqtII ¼ 9.37, p ¼ 0.807), suggesting that the positively charged His-tag is positioned far enough from the pore entrance. Interestingly, when positive voltages (from + 40 to + 100 mV) were applied across the membrane, the current increased nonlinearly (Fig. 4B, inset). When the positive voltages were switched to negative voltages, the channels rap- idly closed and the current dropped close to zero (Fig. 4A). The pores reversibly opened again when a positive voltage was again established. This behavior could be easily explained by the His-tag being on the trans side. At this pH, histidines should not carry an excess of positive charge; however, there is an addi- tional arginine, which is part of a thrombin cleavage site in the spacer (Fig. 1C) and possesses a positive charge, which contributes to the observed behavior. Therefore, an applied negative potential forces the entire N-terminus with the His-tag and linker contain- ing arginine to become closer to the trans entrance of the pore, to enter the pore lumen and to clog it. The rate of closures was voltage-dependent and increased with the magnitude of the negative applied voltage. Furthermore, when the pH of the buffer was lowered to 5.5, i.e. below the pK A of His (pK A ¼ 6.04), which A B C Fig. 2. Formation of pores in PLMs by the wild-type EqtII and mutants. (A) PLMs were composed of 1,2-diphytanoyl-sn-glycero- phosphocholine and 20% (w ⁄ w) SM. The wild-type EqtII and D5 were added at a final concentration of 1–5 n M to the cis side, where a constant voltage (+ 40 mV) was applied. The buffer was 10 m M Tris ⁄ HCl and 100 mM KCl (pH 8.0) on both sides. Only the initial few steps of pore formation are presented for the most rep- resentative traces. Transient currents observed in D5 are indicated by arrows. (B) Histograms show the distribution of pore sizes. The numbers of events from eight and 10 independent experiments were 24 and 83 for the wild-type and D5, respectively. (C) Pore for- mation of S1C and its modified forms (S1C-MTSET + , S1C-MTSEA + , S1C-MTSES – ) at a final concentration of 1–5 nM, under the same experimental conditions as in (A). The representative traces of at least two independent experiments are shown. Topology of equinatoxin II N-terminus K. Kristan et al. 542 FEBS Journal 274 (2007) 539–550 ª 2006 The Authors Journal compilation ª 2006 FEBS drastically increases the positive charge of the His-tag [28], the current at negative voltages was actually com- pletely blocked (Fig. 4A). To gain additional insights into the position of the N-terminus across the membrane, MTS reagents were applied to the cis or trans side of S1C preformed pores (Table 1). The variation in selectivity, which has been shown to be the most sensitive parameter, was meas- ured after the addition of MTS reagents (see Fig. 3B for details). When MTS reagents, which are mem- brane-impermeable, were added to the cis side of the membrane, small or no changes in reversal potential (U rev ) were observed (Table 1). Changes occurred only when MTS reagents were added to the trans side of the membrane, suggesting that the reagents had reac- ted with the thiol group of S1C, and modified the pore selectivity, confirming the trans position of the Ser1 residue. Finally, heterobifunctional maleimeide-poly(ethylene glycol)-N-hydroxysuccinimide was used to chemically modify the thiol group of S1C after pores were already formed. Maleimeide-poly(ethylene glycol)-N- hydroxysuccinimide contains a maleimide group that can react with the thiol group and could possibly clog the channel when covalently attached close to the pore entrance. In multichannel recordings, clo- sures could be seen only when maleimeide-poly(ethy- lene glycol)-N-hydroxysuccinimide was added to the trans side. However, the extent of current reduction did not exceed 15–20% of the total current (Fig. 5A,B). When we performed a three-channel recording, the addition of the maleimeide-poly(ethy- lene glycol)-N-hydroxysuccinimide reagent caused stepwise closures only after maleimeide-poly(ethylene glycol)-N-hydroxysuccinimide was added to the trans side of the membrane (Fig. 5C). Discussion The molecular mechanism of actinoporin pore forma- tion has been unraveled in the last few years, with Fig. 3. Properties of the channels formed by the wild-type and mutants in PLMs. (A) Average conductance of the wild-type and mutant pores. The dashed line shows the wild-type’s conductance for comparison. *p < 0.05; n ¼ 16–60, average ± SD. The experi- mental conditions were as in Fig. 2A. The bar corresponding to D5 refers to the normal channel, i.e. the one with higher conductivity in Fig. 2B. (B) Selectivity of pores formed by the wild-type EqtII and mutants. The wild-type and mutants were added at a final con- centration of 1–5 n M to the cis side of the membrane. Initially, both sides were bathed in a symmetric solution of 10 m M Tris ⁄ HCl and 100 m M KCl (pH 8.0). The KCl concentration was increased step- wise only on the trans side, to reach a final KCl concentration of 1 M (10-fold gradient). U rev values were converted to the reported permeability ratio (P + ⁄ P – ) with the Goldman–Hodgkin–Katz equation (Eqn 2). The dashed line shows the wild-type’s selectivity for com- parison. *p < 0.05; n ¼ 3–9, average ± SD. (C) The dependence of the single-channel current on the applied voltage for the wild-type EqtII and mutants. The I ⁄ V characteristics of the pores formed by the wild-type EqtII (filled squares), S1C (filled diamonds), S1C- MTSES – (open triangles), S1C-MTSET + (open inverted triangles) and D5 (open circles) mutants are shown. They were derived from the amplitude of the current steps caused by square voltage pulses in experiments with membranes containing 5–50 channels. Current values were then normalized at + 40 mV to the current flowing through one single channel, as obtained from the histogram in Fig. 3. All conditions were as described in Fig. 2A. Inset: the ratio of currents (I + ⁄ I – ; in absolute values) when positive and negative voltages were applied. n ¼ 3–9, average ± SD. K. Kristan et al. Topology of equinatoxin II N-terminus FEBS Journal 274 (2007) 539–550 ª 2006 The Authors Journal compilation ª 2006 FEBS 543 particular emphasis on the role of the N-terminal region. It was shown that it needs to be flexible [20,21] and that region 10–28 forms an amphipathic a-helix, so far the only recognized structural element of the final pore [23]. The present study provides additional information about the structure and formation of the EqtII pore. The following conclusions can be drawn from the data presented here: (a) the first and third amino acids of EqtII are exposed to the lumen of the pore; (b) the N-terminal part of EqtII extends to the trans side of the membrane in the final pore, i.e. to the other side than the rest of the membrane-bound mole- cule; and (c) the first five amino acids help to stabilize the final pore. The position of the first and third amino acids in the final pore The change in the fixed charge distribution through the pore affects the pore conductance [30]. Therefore, it can be used for determining the residues exposed to the lumen and the structural organization of the pore [24]. In our experiments, the effects on pore properties were clearly shown for chemically modified S1C, for which large changes in cation selectivity were observed upon chemical modification with MTS reagents. In Malovrh et al. [24], modifications with MTS reagents and comparison of selectivity indices for chemically labeled mutants proved to be very useful for the detec- tion of sites exposed to the pore lumen, particularly the ratio of selectivity indices of a mutant that was chemically modified with MTSES – and MTSEA + [(P + ⁄ P À MTSESÀ ) ⁄ (P + ⁄ P À MTSEAþ )]. This ratio is about 1 when the amino acid side chain is not facing the lumen of the pore, as was observed for most of the mutants in region 10–28 of EqtII [24]. For the mutants exposed to the pore lumen, this ratio should be higher, as selectivity increased with the negative charge attached and decreased with the positive charge. This was indeed observed for the residues from the polar face of the amphipathic helix, with Asp10 showing the Fig. 4. The effect of the N-terminal histidine tag on the voltage-gating properties. (A) Proteins were added to the cis side at a final concentra- tion of 2–20 n M, and the current across the membrane was followed. The buffer was 10 m M Tris ⁄ HCl, 100 mM KCl, and 0.1 mM EDTA (pH 8.0), on both sides, except in the lowest trace, where the buffer was 10 m M Mes, 100 mM KCl, and 0.1 mM EDTA (pH 5.5). Other con- ditions used were the same as described in Fig. 2A. The currents when positive and negative voltages were applied are shown. (B) Cur- rent–voltage dependence of His 6 –EqtII. The inset shows the ratio of currents (I + ⁄ I – ; in absolute values) when positive and negative vol- tages were applied. The protein concentration was 5 n M. Table 1. MTS reagents were added to the cis or trans side in order to study the accessibility of Cys1. Experiments were performed in asymmetric conditions, and U rev was monitored from the first 15 s after the addition of MTS reagents until U rev stabilized (see Fig. 3B for experimental details). The typical changes in U rev of 1–3 experi- ments are reported. The concentration of KCl was 100 m M and 1 M in the cis and trans chambers, respectively. The U rev values for the S1C mutant before the addition of MTS reagents were 38.9 mV, 37.6 mV and 39.1 mV for MTSEA + , MTSET + and MTSES – , respect- ively. D(cis) (mV) D(trans) (mV) MTSEA + + 0.3 ) 3.9 MTSET + 0 ) 2.4 MTSES – + 0.3 + 2.0 Topology of equinatoxin II N-terminus K. Kristan et al. 544 FEBS Journal 274 (2007) 539–550 ª 2006 The Authors Journal compilation ª 2006 FEBS highest ratio of almost 10, suggesting it is located at the constriction of the pore [24]. In our case, the ratio (P + ⁄ P À MTSESÀ ) ⁄ (P + ⁄ P À MTSEAþ ) was about 3 for the S1C mutant, thus clearly indicating that this side chain faces the lumen of the pore. This value is approxi- mately the same as the value observed for Asp17 [24], and would place Ser1 at approximately the same posi- tion with regard to the center of the pore. The introduced charges along the ion conductive pathway should also affect the conductance of pores. Variation of the conductance upon addition of positive charge (MTSEA + or MTSET + ) or negative charge (MTSES – ) has been observed for the solution-exposed amino acids of other PFTs [30,31]. The observed chan- ges were explained primarily by the electrostatic nature of these effects, on the assumption that the channel Fig. 5. The effect of maleimeide-poly(ethylene glycol)-N-hydroxysuccinimide on currents of the wild-type and S1C pores. Multichannel recordings of EqtII (left) and S1C (right) in PLMs. The lowest panel in S1C shows a three channel recording. The concentration of the protein was 1 n M, except for oligochannel recording, where it was 200 pM. Maleimeide-poly(ethylene glycol)-N-hydroxysuccinimide at a final concen- tration of 50 l M was added to the cis or trans compartment, as indicated by arrows. The voltage applied was + 40 mV in all experiments. K. Kristan et al. Topology of equinatoxin II N-terminus FEBS Journal 274 (2007) 539–550 ª 2006 The Authors Journal compilation ª 2006 FEBS 545 maintained a fixed structure. In the present study, only slight differences were observed in the conductance of S1C, modified or not with MTS reagents (Fig. 3A). For a possible explanation of these small variations, we must keep in mind that we modified the position at the tip of a very flexible N-terminus [17,18,32] and that EqtII forms toroidal pores [27] in which helices are not rigidly arranged in the final pore. Accordingly, the high standard deviations in conductance values, which are characteristics of EqtII pores, demonstrate that the helix can slightly change position according to circum- stances [24]. As previously noted [24], we confirm here that EqtII conductance is not the most sensitive parameter with which to study the charge distribution along the pore. We therefore propose that the effects of charges on the conductance of EqtII pores observed for MTS-modified S1C and the low-conductance state of the D5 mutant could be ascribed to an electrostatic effect and to a pore structural rearrangement, respect- ively. A strong effect on selectivity, as a result of the change in the net charge, was observed for the D5 mutant. This mutant showed significantly lower cat- ion selectivity than the wild-type, which is consistent with removal of a negative charge from the ion conductive pathway. The data obtained with D5 therefore suggest that Asp3 is exposed to the pore lumen. Similarly, the addition of a positive charge at position 1 decreased the cation selectivity, and the addition of a negative charge increased it. According to Malovrh et al. [24], the modulation of negative charges is crucial for defining the electrophysiologic characteristics of EqtII. In particular, selectivity is the most sensitive electrophysiologic parameter and the one that is most affected by charge modifications in the pore lumen. Furthermore, current dependence on the applied voltage was studied for chemically modified S1C and D5. Deletion of the first five amino acid residues (D5) and positive charge addition (S1C-MTSET + ) had the greatest effect. The strong asymmetry of the I ⁄ V curves of D5 and S1C-MTSET + provides the first indi- cation that the N-terminus is exposed to the trans side of the membrane. Furthermore, values of the I + ⁄ I – ratio larger than 1, as measured for those mutants, strongly suggest a trans position of Ser1 (Fig. 3C, inset). In this case, the local trans concentration of cat- ions is lower than the bulk concentration, due to the repulsive effect of MTSET + ,K + (and Cl – ), so they move from cis to trans (or trans to cis) according to both electrical and concentration gradients, leading to a higher current. The opposite happens when a negat- ive voltage is applied. The position of the N-terminus in the final pore The asymmetry of the I ⁄ V curves indicates that the first and third positions are exposed to the lumen of the pore. Additional insights into the position of the N-terminus were obtained with the His-tagged version of EqtII. The data showed that His 6 –EqtII forms channels of lower conductance, has an asymmetric I ⁄ V curve, and exhibits rapid closures of pores at a negat- ive applied potential voltage. As we did not observe any changes in the selectivity of the pores formed by His 6 –EqtII, the changes in current must again be due to changes in the pore structure and flexibility of the His-tag. The above observations can be interpreted to mean that the His-tag is translocated to the trans side of the membrane and then blocks the channel when negative voltages are applied. This is likely, because the His-tag with the linker possesses at least one posit- ive charge at the pH of the buffer used (pH 8), and can therefore act as a voltage-dependent gate. At pH 5.5 (Fig. 4), most of the histidines are protonated and thus the N-terminus of His 6 –EqtII carries a large excess of positive charge. Consistently, the rates of clo- sure and opening of the pore, as well as the blocking efficiency at negative voltages, are drastically increased. The most likely mechanism by which it may close the pore is by inserting into the pore lumen and thus obstructing the ion permeability when a negative potential is applied. This mechanism would be analog- ous to the ‘ball and chain’ mechanism of channel inac- tivation [33]. The voltage-dependent closures of His 6 – EqtII pores are also extremely similar to those observed for His-tagged diphtheria toxin [33]. In that study, it was shown that diphtheria toxin is able to translocate the His-tag at the N-terminal region across the lipid membrane. The same occurs in the case of EqtII; how- ever, the pore formation efficiency of His 6 –EqtII is reduced, as this mutant is less hemolytically active. A similar reduction in permeabilizing activity when the His-tag was present was reported for the homologs magnificalysin [34] and sticholysin II [35]. In fact, the N-terminus is critical for the permeabilizing activity of EqtII, and longer tags rendered it inactive [20]. Topological experiments with thiol-modifiable rea- gents were also performed. MTS reagents were added to either side of the membrane when S1C pores were already formed, but changes were observed only upon addition to the trans side of the membrane (Table 1). Finally, maleimeide-poly(ethylene glycol)-N-hydroxy- succinimide only had observable effects when added to the trans side (Fig. 5C). Altogether, the data obtained indicate that the N-terminal part of EqtII is exposed to the trans side when the transmembrane pore is formed. Topology of equinatoxin II N-terminus K. Kristan et al. 546 FEBS Journal 274 (2007) 539–550 ª 2006 The Authors Journal compilation ª 2006 FEBS Stabilization of pores by first five amino acids The most unusual behavior was observed with D5 (Fig. 2). Unstable pores with half the conductance of the fully open D5 pores were observed. The current data cannot exclude the possibility that the smaller currents are due to the formation of channels by only three or even fewer monomers. In this case, the N-terminal end would help to assemble monomers in the final pore. However, we believe that these smaller channels represent an intermediate on the way to the final pore, where the lack of five amino acids pre- vents stabilization of final fully open pores. In this model, the first five amino acids would act as an anchor, which would help to restrict the N-terminus to the trans side of the membrane. Region 1–5 is highly hydrophobic in actinoporins (Fig. 1D). The Asp present at position 3 in EqtII is an exception, as most actinoporins possess a hydrophobic amino acid at that position (Fig. 1D). The fourth amino acid is a bulkier hydrophobic Val or Leu and could actually have the most important role. The first five amino acids are also highly flexible and were not resolved in the crystal structure of EqtII [17]. Recently, an NMR structure of a peptide corresponding to region 1–32 was determined in the presence of dodecylphospho- choline micelles. It formed a continuous helix from residues 6 to 28, but again the first five amino acids showed the highest flexibility [32]. This region would have a similar role as in aerolysin, a b-pore-forming toxin from bacteria, where hydrophobic amino acids from the tip of the b-loop anchor the b-barrel in the membrane [36]. Experimental procedures Materials Bovine brain SM and 1,2-diphytanoyl-sn-glycerophospho- choline were obtained from Avanti Polar Lipids (Alabaster, AL). All other materials were from Sigma (Milan, Italy), unless stated otherwise. Cloning, expression and isolation of the mutants The construction of expression vectors of mutants S1C and D5 (deleted first five amino acids of EqtII) has been des- cribed previously (Fig. 1C) [22,29]. The wild-type EqtII, S1C and D5 were expressed in an E. coli BL21 (DE3) strain and purified from the bacterial cytoplasm as described else- where [37]. The wild-type EqtII was also constructed as a His 6 fusion protein, which contains an N-terminal hexa-his- tidine tag and the thrombin cleavage site (Fig. 1C). His 6 – EqtII was expressed in an E. coli BL21 (DE3) pLysS strain and purified from the bacterial cytoplasm by Ni-chelate chromatography [38]. All mutants, fusion proteins and the wild-type were purified to homogeneity on SDS ⁄ PAGE gels. Hemolytic activity Hemolytic activity was measured by the use of a microplate reader (MRX; Dynex Technologies, Deckendorf, Ger- many). A suspension of bovine red blood cells was pre- pared in hemolysis buffer (0.13 m NaCl, 0.02 m Tris ⁄ HCl, pH 7.4) from well-washed erythrocytes. One hundred microliters of erythrocyte suspension with A 630 ¼ 0.5 was added to 100 lL of two-fold serially diluted proteins. Hemolysis was then monitored turbidimetrically by measur- ing the absorbance at 630 nm for 20 min at room tempera- ture. The results are presented as c 50 , which is the concentration of a protein that produces 50% of the max- imal rate of hemolysis. Chemical modification using MTS derivatives Mutant S1C was chemically modified with MTS reagents to introduce either a positive or a negative charge at the thiol group [30,39]. MTSEA + and MTSET + were used for the introduction of positive charges, and MTSES – was used to introduce a negative charge (all from Biotium, Inc., Fremont, CA). S1C at a concentration of 10– 50 lgÆmL )1 (0.5–2 lm) in water was preincubated over- night in a 200 molar excess of dithiothreitol (0.1–0.4 mm). MTS reagents, freshly dissolved in water, were then added at 1000 molar excess (0.5–2 mm). After 1 h of incubation at room temperature, after which the majority of the rea- gent had been hydrolyzed according to the manufacturer’s specifications, the modified samples were used for PLM experiments. The final concentrations of MTS reagents or dithiothreitol in the cis chamber after addition of the sample to the PLM were below 5 lm. The pore propert- ies of the wild-type EqtII were not affected by the MTS reagents [24]. PLM experiments Solvent-free PLMs were composed of 1,2-diphytanoyl- sn-glycerophosphocholine and 20% SM (w ⁄ w) [40]. The chambers were made of Teflon, and the volume of the chambers was 2 mL. The septum between the chambers was also made of Teflon and contained a 100 lm hole. The protein was added at nanomolar concentrations to stable, preformed bilayers on the cis side only (the cis side is where the electrical potential was applied, and the trans side was grounded). All experiments were started in symmetric con- ditions, using a buffer comprising 10 mm Tris ⁄ HCl and 100 mm KCl (pH 8.0) on both sides of the membrane. For K. Kristan et al. Topology of equinatoxin II N-terminus FEBS Journal 274 (2007) 539–550 ª 2006 The Authors Journal compilation ª 2006 FEBS 547 experiments with His 6 –EqtII fusion protein, 0.1 mm EDTA was included in the buffer. A defined voltage, generally + 40 mV, was applied across the membrane. Miniature magnetic stir bars stirred the solutions on both sides of the membrane. The currents across the bilayer were measured, and the conductance (G) was determined as follows [41]: GðpSÞ¼IðpAÞ=U ðVÞð1Þ where I is the current through the membrane, and U is the applied transmembrane potential. Macroscopic currents were recorded by a patch clamp amplifier (Axopatch 200, Axon Instruments, Foster City, CA). A PC equipped with a DigiData 1200 A ⁄ D converter (Axon Instruments) was used for data acquisition. The current traces were filtered at 100 Hz and acquired at 500 Hz by the computer using axoscope 8 software (Axon Instruments). All measurements were performed at room temperature. For the selectivity measurement, KCl concentration was increased stepwise on the trans side only to finally form a 10-fold gradient. At each concentration, the potential neces- sary to zero the transmembrane current (i.e. the reversal potential U rev ), was determined. From the reversal poten- tial, the ratio of the cation over anion permeability (P + ⁄ P – ) was calculated using the Goldman–Hodgkin–Katz equation [42–44]: P þ =P À ¼½ða trans =a cis Þ expðeU rev =kTÞÀ1= ½ða trans =a cis ÞÀexpðeU rev =kTÞ ð2Þ where a trans and a cis are the activities of KCl on the trans side and the cis side, respectively [a, thermodynamic activity on trans or cis side of membrane; U rev , reversal potential; e, elementary charge; k, Boltzmann constant; T, absolute temperature (at 23 °C kT/e ¼ 25 mV)]. kT ⁄ e is % 25 mV at room temperature. The P + ⁄ P – values repor- ted were measured at the same conditions, which were 100 mm KCl in the cis chamber and 1 m KCl in the trans chamber. For topological experiments with the S1C mutant, the protein was preincubated for 30 min with 20 mm dithio- threitol and added to PLMs to allow pore formation. MTS reagents or heterobifunctional maleimeide-poly(ethylene glycol)-N-hydroxysuccinimide (molecular mass ¼ 3400 Da; Nektar Therapeutics, Huntsville, AL) were then added to the cis or trans solution. The final concentrations used were 1or2mm for MTS reagents and 50 lm for maleimeide- poly(ethylene glycol)-N-hydroxysuccinimide. Acknowledgements The Slovenian authors were supported by grants from the Slovenian Research Agency (Ljubjana, Slovenia). GV was supported by fellowships from the CNR Insti- tute of Biophysics (Trento, Italy). References 1 Gouaux E (1997) Channel-forming toxins: tales of transformation. Curr Opin Struct Biol 7, 566–573. 2 Parker MW & Feil SC (2005) Pore-forming protein tox- ins: from structure to function. Prog Biophys Mol Biol 88, 91–142. 3 Tweten RK (2005) Cholesterol-dependent cytolysins, a family of versatile pore-forming toxins. Infect Immun 73, 6199–6209. 4 Heuck AP, Tweten RK & Johnson AE (2001) b-Barrel pore-forming toxins: intriguing dimorphic proteins. Bio- chemistry 40, 9065–9073. 5 Neumeyer T, Tonello F, Molin FD, Schiffler B & Benz R (2006) Anthrax edema factor, voltage-depen- dent binding to the protective antigen ion channel and comparison to LF binding. J Biol Chem 281, 32335–32343. 6 Anderluh G & Lakey JH (2005) Lipid interactions of a-helical protein toxins. In Protein–Lipid Interactions. From Membrane Domains to Cellular Networks (Tamm LK, ed.), pp. 141–162. Wiley-VCH, Weinheim. 7 Lakey JH & Slatin SL (2001) Pore-forming colicins and their relatives. Pore-Forming Toxins 257, 131–161. 8 Anderluh G & Mac ˘ ek P (2002) Cytolytic peptide and protein toxins from sea anemones (Anthozoa: Actini- aria). Toxicon 40 , 111–124. 9 Van der Goot FG, Gonza ´ lez-Man ˜ as JM, Lakey JH & Pattus F (1991) A ’molten-globule’ membrane-insertion intermediate of the pore-forming domain of colicin A. Nature 354, 408–410. 10 Song LZ, Hobaugh MR, Shustak C, Cheley S, Bayley H & Gouaux JE (1996) Structure of staphylococcal a-hemolysin, a heptameric transmembrane pore. Science 274, 1859–1866. 11 Viero G, Cunaccia R, Pre ´ vost G, Werner S, Monteil H, Keller D, Joubert O, Menestrina G & Dalla Serra M (2006) Homologous versus heterologous interactions in the bicomponent staphylococcal gamma-haemolysin pore. Biochem J 394, 217–225. 12 Panchal RG, Smart ML, Bowser DN, Williams DA & Petrou S (2002) Pore-forming proteins and their application in biotechnology. Curr Pharm Biotechnol 3, 99–115. 13 Potrich C, Tomazzolli R, Dalla Serra M, Anderluh G, Malovrh P, Mac ˘ ek P, Menestrina G & Tejuca M (2005) Cytotoxic activity of a tumor protease-activated pore- forming toxin. Bioconjug Chem 16, 369–376. 14 Astier Y, Bayley H & Howorka S (2005) Protein com- ponents for nanodevices. Curr Opin Chem Biol 9, 576– 584. 15 Belmonte G, Pederzolli C, Mac ˘ ek P & Menestrina G (1993) Pore formation by the sea anemone cytolysin equinatoxin II in red blood cells and model lipid mem- branes. J Membrane Biol 131, 11–22. Topology of equinatoxin II N-terminus K. Kristan et al. 548 FEBS Journal 274 (2007) 539–550 ª 2006 The Authors Journal compilation ª 2006 FEBS [...]...K Kristan et al 16 Tejuca M, Dalla Serra M, Ferreras M, Lanio ME & Menestrina G (1996) Mechanism of membrane permeabilisation by sticholysin I, a cytolysin isolated from the venom of the sea anemone Stichodactyla helianthus Biochemistry 35, 14947–14957 17 Athanasiadis A, Anderluh G, Macek P & Turk D ˘ (2001) Crystal structure of the soluble form of equinatoxin II, a pore- forming toxin from the sea... MG, Zhang W, Anderluh G, Hansen PE & Norton RS (2002) Solution structure of the eukaryotic poreforming cytolysin equinatoxin II: implications for pore formation J Mol Biol 315, 1219–1229 19 Mancheno JM, Martin-Benito J, Martinez-Ripoll M, Gavilanes JG & Hermoso JA (2003) Crystal and electron microscopy structures of sticholysin II actinoporin reveal insights into the mechanism of membrane pore formation... Ion and nonelectrolyte permeability properties of channels formed in planar lipid bilayer membranes by the cytolytic toxin from the sea anemone, Stoichactis helianthus J Membrane Biol 55, 203–211 26 Valcarcel CA, Dalla Serra M, Potrich C, Bernhart I, Tejuca M, Martinez D, Pazos F, Lanio ME & Menestrina G (2001) Effects of lipid composition on membrane permeabilization by sticholysin I and II, two Topology... JK, Frisanco M, Guella G, Dalla Serra M, Anderluh G, Separovic F & Norton RS (2006) Structure and activity of the N-terminal region of the eukaryotic cytolysin equinatoxin II Biochemistry 45, 1818–1828 Senzel L, Huynh PD, Jakes KS, Collier RJ & Finkelstein A (1998) The diphtheria toxin channel-forming T domain translocates its own NH2-terminal region across planar bilayers J Gen Physiol 112, 317–324... formation by equi˘ natoxin, a eukaryotic pore- forming toxin, requires a flexible N-terminal region and a stable beta-sandwich J Biol Chem 279, 46509–46517 22 Anderluh G, Barlic A, Podlesek Z, Macek P, Pungercar ˘ ˘ ˘ J, Gubensˇ ek F, Zecchini ML, Dalla Serra M & Menestrina G (1999) Cysteine-scanning mutagenesis of an eukaryotic pore- forming toxin from sea anemone ) topology in lipid membranes Eur J Biochem... G & Anderluh G (2000) Structure–function studies of tryptophan mutants of equinatoxin II, a sea anemone pore- forming protein Biochem J 346, 223–232 24 Malovrh P, Viero G, Dalla Serra M, Podlesek Z, Lakey JH, Macek P, Menestrina G & Anderluh G (2003) A ˘ novel mechanism of pore formation ) membrane penetration by the N-terminal amphipathic region of equinatoxin J Biol Chem 278, 22678–22685 25 Varanda... permeabilization by sticholysin I and II, two Topology of equinatoxin II N-terminus 27 28 29 30 31 32 33 34 35 36 37 38 cytolysins of the sea anemone Stichodactyla helianthus Biophys J 80, 2761–2774 Anderluh G, Dalla Serra M, Viero G, Guella G, Macek ˘ P & Menestrina G (2003) Pore formation by equinatoxin II, a eukaryotic protein toxin, occurs by induction of nonlamellar lipid structures J Biol Chem 278, 45216– 45223... Khoo HE (2000) A new cytolysin from the sea anemone, Heteractis magnifica: isolation, cDNA cloning and functional expression Biochim Biophys Acta 1478, 9–18 Pazos IF, Martinez D, Tejuca M, Valle A, del Pozo A, Alvarez C, Lanio ME & Lissi EA (2003) Comparison of pore- forming ability in membranes of a native and a recombinant variant of sticholysin II from Stichodactyla helianthus Toxicon 42, 571–578 Iacovache... pore- forming toxins EMBO J 25, 457–466 ˇ Anderluh G, Pungercar J, Strukelj B, Macek P & ˘ ˘ Gubensˇ ek F (1996) Cloning, sequencing and expression of equinatoxin II Biochem Biophys Res Commun 220, 437–442 Anderluh G, Gokce I & Lakey JH (2003) Expression of proteins using the third domain of the Escherichia coli FEBS Journal 274 (2007) 539–550 ª 2006 The Authors Journal compilation ª 2006 FEBS 549 Topology... Kristan K, Podlesek Z, Macek P, Turk D, Gonzalez˘ Manas JM, Lakey JH et al (2002) Two-step mem˜ brane binding by equinatoxin II, a pore- forming toxin from the sea anemone, involves an exposed aromatic cluster and a flexible helix J Biol Chem 277, 41916– 41924 ´ 21 Kristan K, Podlesek Z, Hojnik V, Gutierrez-Aguirre I, ´ Guncar G, Turk D, Gonzalez-Manas JM, Lakey JH, ˘ ˜ Macek P & Anderluh G (2004) Pore . The equinatoxin N-terminus is transferred across planar lipid membranes and helps to stabilize the transmembrane pore Katarina Kristan 1 , Gabriella. stabilize the toroidal pore [19]. Thus, the purpose of this work was to gain further insight into the structure of the EqtII pore, especially the topology of the