Mitochondria Permeabilization by a Novel Polycation Peptide BTM-P1*

Bacillus thuringiensis subsp. medellin is known to produce the Cry11Bb protein of 94 kDa, which is toxic for mosquito larvae due to permeabilization of the plasma membrane of midgut epithelial cells. Earlier we found that a 2.8-kDa novel peptide BTM-P1, which was artificially synthesized taking into account the primary structure of Cry11Bb endotoxin, is active against several species of bacteria. In this work we show that BTM-P1 induces cyclosporin A-insensitive swelling of rat liver mitochondria in various salt solutions but not in the sucrose medium. Inorganic phosphate and Ca2+ significantly increased this effect of the peptide. The uncoupling action of BTM-P1 on oxidative phosphorylation was stronger in the potassium-containing media and correlated with a decrease of the inner membrane potential of mitochondria. In isotonic KNO3, KCl, or NH4NO3 media, a complete drop of the inner membrane potential was observed at 1–2 μg/ml of the peptide. The peptide-induced swelling was increased by energization of mitochondria in the potassium-containing media, but it was inhibited in the NaNO3, NH4NO3, and Tris-NO3 media. All mitochondrial effects of the peptide were completely prevented by adding a single N-terminal tryptophan residue to the peptide sequence. We suggest a mechanism of membrane permeabilization that includes a transmembrane- and surface potential-dependent insertion of the polycation peptide into the lipid bilayer and its oligomerization leading to formation of ion channels and also to the mitochondrial permeability transition pore opening in a cyclosporin A-insensitive manner.

Mitochondrial voltage-dependent anion channel and its interaction with other proteins seem to be directly related to the metabolic specificity and permeabilization resistance of mitochondria in tumor cells. It has been reported that oncogenic C-Raf kinase inhibits cytochrome c release from mitochondria in growth factor-starved cells, and it has been hypothesized that the voltage-dependent anion channel mediates the switch to aerobic glycolysis (29). This hypothesis is consistent with theoretical models predicting a higher probability of the Crabtree effect in the case of decreased permeability of the mitochondrial voltage-dependent anion channel and a high amount of hexokinase II attached to the outer mitochondrial membrane (30,31). Lithium chloride, which has an anti-proliferative effect on melanoma cells (32), induces detachment of hexokinase from mitochondria (33), and vice versa, the interaction of hexokinase II with voltage-dependent anion channel inhibits the release of cytochrome c from mitochondria induced by pro-apoptotic protein Bax (34). In this sense some synthetic polycation peptides are also able to inhibit the mitochondrial permeability transition (35).
Understanding the mechanisms of action of some types of proteins and peptides on the structural and functional state of mitochondria seems to be important for the development of apoptosis-regulating technologies, particularly of the anticancer therapy, and for the design of new classes of antibiotics, taking into account a certain similarity between mitochondria and bacteria. Among the positively charged peptides, mastoparan, a 14-amino acid amphipathic peptide from wasp venom, has been found to cause mitochondrial permeabilization (36). Mastoparan has been shown to take the ␣-helical conformation upon binding with phospholipid membrane (37) and to cause liposome permeabilization in a transmembrane potential-dependent manner (36) that seems to be the mechanism for mitochondria as well (38). Recently, it has been shown that the liposome permeation ability of eumenine mastoparan, a novel natural tetradecapeptide with a broad-spectrum inhibitory activity against Gram-positive and Gram-negative bacteria, is correlated with the amount of helical conformation in the membrane (39).
Most of the signal sequences of mitochondrial proteins with a net positive charge have been shown to induce potentialdependent swelling of mitochondria and to activate the mitochondrial multiple conductance channel (38,40) that has been suggested to be distinct from both the classical permeability transition pore and the mastoparan-induced permeability increase (40). Many membrane-active peptides have been reported to function as a defense system against bacteria in plants and animals (41)(42)(43). Among them, the positively charged peptides magainins, secreted by the skin of the am-phibia Xenopus laevis, uncouple oxidative phosphorylation and significantly inhibit respiration of rat liver mitochondria and display antimicrobial (41) and anticancer activity (21,22). Not only plants and animals but some bacteria also produce toxic peptides and protoxins against neighboring organisms. Bacillus thuringiensis subsp. medellin, for instance, produce crystal proteins, particularly Cry11Bb, that is toxic for mosquito larvae (44). At least some types of Cry family proteins are activated by the midgut proteases to yield a membrane-active polypeptide(s) (45). A non-insecticidal and non-hemolytic B. thuringiensis crystal protein selectively kills human cancer cells in culture by non-apoptotic and apoptotic mechanisms (46).
Here we demonstrate various effects of a new polycation peptide (BTM-P1) 1 on structural and functional state of rat liver mitochondria. The peptide is composed of 26 amino acid residues and derived from the primary structure of Cry11Bb protein of B. thuringiensis subsp. medellin. Its antibacterial activity has been recently demonstrated and related to the membrane ionic permeabilization. 2 In the present work we show that BTM-P1 in low concentrations uncouples respiration and oxidative phosphorylation of rat liver mitochondria, causes the drop of the inner membrane potential, and induces mitochondrial swelling in various salt media. The permeabilization effect of the peptide is potential-dependent, increases with an increase in potassium concentration, and in general depends on the ionic strength and ionic composition of incubation medium. Permeabilizing activity of the peptide can be completely prevented by an additional N-terminal tryptophan residue (the peptide BTM-WP1 of 27 amino acid residues) that allows the suggestion of a possible mechanism of the potential-dependent permeabilization of biomembranes by BTM-P1 and other similar peptides.

EXPERIMENTAL PROCEDURES
Materials-Sucrose, D-mannitol, Hepes, Trizma base (Tris base), bovine serum albumin (free of fatty acids, fraction V), cyclosporin A, rotenone, carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP), phosphoric and succinic acids, ADP, KCl, and other salts were purchased from Sigma. All chemicals were of analytical grade. The peptides BTM-P1 and BTM-WP1 were synthesized by Fundación Instituto de Inmunología de Colombia (Bogota, Colombia) by the method of solid phase synthesis, according to the amino acid sequences proposed by some of the authors of this work. The peptides were purified by high pressure liquid chromatography, analyzed by mass spectrometry (matrix-assisted laser desorption ionization time-of-flight), lyophilized, and kept in powder at 4°C until its use. 2 Isolation of Mitochondria-White male rats of 200 -250 g starved for 12-14 h were used for isolation of liver mitochondria by the standard procedure as described earlier (48) with some modifications. Mitochondria were finally suspended in the medium composed of 210 mM mannitol, 70 mM sucrose, 20 M EGTA, 10 mM Hepes-Tris, pH 7.2, with 0.3 mg/ml bovine serum albumin (free of fatty acids). Mitochondrial protein concentration was evaluated by the fast volumetric test (48) and finally determined by the biuret method after finishing all experiments.
Respiration and Oxidative Phosphorylation-Respiration and oxidative phosphorylation of mitochondria were measured polarographically using the modified Cole-Parmer oxygen electrode (49) and the Cole-Parmer oxygen meter connected to the Linseis L250-E recorder. Mitochondria were added at concentration of 1 mg of protein/ml to the incubation medium containing 150 mM sucrose, 50 mM KCl, 50 M EGTA, 10 mM Hepes-Tris buffer, pH 7.2 (sucrose-KCl medium), or to the medium containing 250 mM sucrose, 50 M EGTA, 10 mM Hepes-Tris buffer, pH 7.2 (sucrose medium), both supplemented with 5 mM succinate-Tris and 5 mM phosphate-Tris, pH 7.2. After the addition of 2.5 M rotenone and registration of respiration in the metabolic state 4, 270 M ADP was added to initiate oxidative phosphorylation. The initial concentration of oxygen in the media has been considered to be equal to 200 nmol/ml, according to the atmospheric pressure in Medellín. The respiratory control ratio, calculated as the ratio of the oxygen consumption rates after and before ADP addition, was higher than six.
Mitochondrial Swelling-Mitochondrial swelling was monitored by apparent light absorbance at 540 nm, A 540 , using a SP-850 Turner spectrophotometer equipped additionally with a magnetic stirrer and a thermostatic chamber. The apparent absorbance curve was recorded with the Linseis L250-E recorder connected to the spectrophotometer. Mitochondria (0.5 mg of protein/ml), 2.5 M rotenone, and where indicated 0.5 M cyclosporin A were added to the following incubation media, all containing 10 mM Hepes-Tris buffer, pH 7.2, and 50 M EGTA in addition to 250 mM sucrose (sucrose medium), 150 mM sucrose and 50 mM KCl (sucrose-KCl medium), 125 mM KCl (KCl medium), 125 mM KNO 3 (KNO 3 medium), 125 mM NH 4 NO 3 (NH 4 NO 3 medium), 125 mM NaNO 3 (NaNO 3 medium), or 125 mM Tris-NO 3 (Tris-NO 3 medium) with or without 5 mM succinate-Tris, pH 7.2. Peptides were added as aqueous solutions.
Membrane Potential-The inner membrane electrical potential of mitochondria was monitored with the potential-sensitive fluorescent probe safranin O (520 nm excitation, 580 nm fluorescence) at a concentration of 10 M, as described in Wieckowski and Wojtczak (50), using the sucrose, sucrose-KCl, KCl, NH 4 NO 3 , and KNO 3 media (see above) supplemented with 5 mM succinate-Tris, pH 7.2, with or without 5 mM phosphate-Tris, pH 7.2. In the case of energization of mitochondria with 1 mM ATP, the sucrose-KCl and KCl media (see above) were used without the addition of succinate and phosphate. Mitochondria were added at a concentration of 0.5 mg protein/ml together with 2.5 M rotenone. All samples were constantly stirred with a magnetic stirrer and maintained at 30°C. Typical curves of at least three independent experiments are presented in figures.

RESULTS
The obtained data show that BTM-P1, at a concentration of 4 g/ml in the sucrose-KCl medium, increased the state 4 respiration of rat liver mitochondria almost up to the rate observed for the metabolic state 3 respiration (Fig. 1, a-c). Uncoupling activity of the peptide (Fig. 1c) was slightly less than that observed for 0.1 mM DNP (Fig. 1d). In the range of 1-4 g/ml, the peptide stimulated mitochondrial respiration in a dose-dependent manner (Fig. 1b). Higher concentrations, 5 g/ml or above, caused some inhibitory effect (data not shown).
The uncoupling activity of the peptide was significantly lower in the sucrose medium ( Fig. 1, e and f) in comparison with that in the sucrose-KCl medium ( Fig. 1, b and c). It was possible to observe ATP synthesis with ADP/O ϭ 1.44 (1.87 in the control) even in the presence of the peptide at a concentration of 2 g/ml in the sucrose medium ( Fig. 1e), although the respiratory control ratio was significantly decreased (3.2 in comparison to 6.9 in the control). In the presence of 4 g/ml BTM-P1 in the sucrose medium, i.e. at 4 g per 1 mg of mitochondrial protein, the respiration rate may be further stimulated by 0.1 mM DNP (Fig. 1f) up to the value observed in the presence of DNP without the peptide (Fig. 1g).
The effect of BTM-P1 on the inner membrane potential increased with an increase of concentration of KCl in isotonic incubation media (Fig. 2, A-C). Only a slight decrease of the inner membrane potential was caused by 1 g/ml BTM-P1 (2 g per 1 mg of mitochondrial protein) in the sucrose medium with or without inorganic phosphate ( Fig. 2A), but almost complete deenergization occurred in the KCl medium (Fig. 2C). A higher effect of the peptide was observed in the presence of inorganic phosphate in the KCl medium, where even 1 g/ml of BTM-P1 caused biphasic and a complete drop of the inner membrane potential (Fig. 2C, b). In the sucrose medium the effect of BTM-P1 was also higher in the presence than in the absence of inorganic phosphate, but only after a third addition of the peptide, up to final concentration of 3 g/ml ( Fig. 2A, b).
Mitochondria were characterized by a higher value of the inner membrane potential in the presence of 5 mM inorganic 1 The abbreviations used are: BTM-P1, the peptide VAPIAKYLATA-LAKWALKQGFAKLKS; BTM-WP1, the peptide WVAPIAKYLATAL AK-WALKQGFAKLKS; FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone; DNP, 2,4-dinitrophenol. 2  (curve b), the addition of inorganic phosphate to the energized mitochondria in the sucrose-KCl medium caused a further increase of the inner membrane potential that resulted from a decrease of the transmembrane ⌬pH caused by ⌬pH-dependent phosphate transport with subsequent recovery of the proton electrochemical gradient by the respiratory chain function, mainly due to an increase in the transmembrane electrical potential. Without inorganic phosphate, the effect of 1 g/ml BTM-P1 on the inner membrane potential of mitochondria in the sucrose-KCl medium was only slightly less than in the KCl-medium (Fig. 2, B, a, and C, a, respectively).
Significantly different, biphasic kinetics of the potential drop were observed in these media with inorganic phosphate after the addition of the peptide; the first phase of depolarization of the inner mitochondrial membrane in the KCl medium was followed by a second phase of further autocatalytically increased membrane permeabilization (Fig. 2C, b). In the sucrose-KCl medium, in contrast, the first phase of depolarization was changed by a repolarization process (Fig. 2B, b) that looks like an excitable membrane behavior after under-threshold stimulus.
A depolarization-repolarization process was also observed in the NH 4 NO 3 medium with inorganic phosphate; a small decrease of the inner membrane potential was observed just after the addition of 1 g/ml BTM-P1, but very soon its value was almost completely recovered (Fig. 2D, b). A higher depolariza-tion effect, with a subsequent tendency to recover the inner membrane potential, occurred in the medium without phosphate ( Fig. 2D, a). Although partial (Fig. 2B, b) or complete recovery (Fig. 2D, b) of the inner membrane potential was observed after the first addition of 1 g/ml of BTM-P1, the second addition of 1 g/ml of the peptide (i.e. to the final concentration of 4 g per 1 mg of mitochondrial protein) caused fast and complete depolarization of the membrane.
The inner membrane potential generated by ATP hydrolysis was also altered by the peptide, with almost complete depolarization after the second addition of 1 g/ml BTM-P1. A somewhat higher sensitivity of mitochondria to the peptide was observed in the KCl medium (Fig. 2E, c) than in the sucrose-KCl medium (Fig. 2E, d).
To evaluate the character and the extent of permeabilization effect of BTM-P1 on the inner membrane of mitochondria, the mitochondrial swelling was studied in the media composed of sucrose or various salts buffered with 10 mM Hepes-Tris, pH 7.2, and supplemented with 50 M EGTA to decrease a probability of calcium-and phosphate-induced permeability transition. The obtained data show that 4 g/ml of BTM-P1 had no appreciable induction of the swelling of energized or non-energized mitochondria in the sucrose medium (Fig. 3, Sucrose, curves b and a, respectively). A small induction, higher in the case of energized mitochondria, was observed in the sucrose-KCl medium (Fig. 3, Sucrose-KCl). A significant effect of the peptide was observed in the KCl medium, and the swelling of high amplitude occurred in the KNO 3 medium, with the rate of swelling significantly higher for energized than non-energized mitochondria (Fig. 3, KCl and KNO 3 , curves b and a, respectively). The difference in the rate and the extent of the swelling in the KCl and KNO 3 media may be related to the known higher permeability of the inner membrane to NO 3 Ϫ than to Cl Ϫ ions (1). Cyclosporin A did not influence mitochondrial swelling induced by BTM-P1 in the potassium containing media, as shown in Fig. 3 for the KNO 3 medium.
An opposite effect of the inner membrane energization on mitochondrial swelling, induced by BTM-P1, was obtained using NH 4 NO 3 , NaNO 3 , and Tris-NO 3 incubation media (Fig. 3). Fast swelling of high amplitude was induced by the peptide in the case of non-energized but not energized mitochondria in the NH 4 NO 3 medium (Fig. 3, NH 4 NO 3 , a and b, respectively). The effect of the peptide was significantly higher than that caused by a completely uncoupling concentration of FCCP under the same conditions (Fig. 3, NH 4 NO 3 , c). Cyclosporin A diminished but not prevented the swelling of non-energized mitochondria in the NH 4 NO 3 medium under the influence of BTM-P1 (Fig. 3, NH 4 NO 3 , ϩCsA). High amplitude swelling of non-energized mitochondria induced by the peptide was also observed in the NaNO 3 medium (Fig. 3, NaNO 3 , a), and a minor effect was registered in the Tris-NO 3 medium (Fig. 3, Tris-NO 3 , a).
BTM-P1 did not induce swelling of energized mitochondria in the NH 4 NO 3 medium (Fig. 3, NH 4 NO 3 , b) at least during 4 -5 min of incubation without or with cyclosporin A, except an initial swelling of low amplitude. The peptide induced a fast semi-maximal swelling of energized mitochondria in the NaNO 3 medium (Fig. 3, NaNO 3 , b) as well as less profound biphasic swelling in the Tris-NO 3 medium (Fig. 3, Tris-NO 3 , b).
The data in Fig. 4A show that Ca 2ϩ induces mitochondrial swelling in the sucrose (curve a) or sucrose-KCl (curve d) media with inorganic phosphate. BTM-P1, which was added before Ca 2ϩ , after phosphate (Fig. 4A, b), or before both Ca 2ϩ and phosphate (Fig. 4A, c) in the sucrose medium or after phosphate and Ca 2ϩ in the sucrose-KCl medium (Fig. 4 A, e) significantly accelerated swelling of energized mitochondria. At a concentration of 2 g/ml in the sucrose-KCl medium with inorganic phosphate, BTM-P1 induced swelling of energized (Fig.  4A, g) but not deenergized mitochondria (Fig. 4A, f). Note, in the sucrose-KCl medium without phosphate, the peptide, even at a concentration of 4 g/ml did not cause a significant swelling of energized or non-energized mitochondria (Fig. 3, Sucrose-KCl).
The swelling of energized mitochondria may be also initiated by inorganic phosphate added after Ca 2ϩ (Fig. 4B, a). FCCP, added to mitochondria before Ca 2ϩ and phosphate, completely prevented not only Ca 2ϩ -and phosphate-dependent swelling but also the swelling of high amplitude induced by the subsequent addition of 4 g/ml BTM-P1 (Fig. 4B, c). In contrast to FCCP, BTM-P1, added at 2 g/ml before Ca 2ϩ and phosphate, significantly accelerated the Ca 2ϩ -and phosphate-induced swelling (Fig. 4B, b), in comparison to the control without the peptide (Fig. 4B, a). FCCP, added together with mitochondria in the BTM-P1-containing medium, was able to completely prevent this type of accelerated swelling (Fig. 4B, d).
The presence of phosphate in the KNO 3 incubation medium increased the sensitivity of mitochondria to a small concentration of BTM-P1, 0.5 g/ml (0.18 M), resulting in a biphasic mitochondrial swelling (Fig. 4B, f) in comparison to the control without phosphate (Fig. 4B, e). The first phase of swelling seems to be due to direct permeabilization of the inner mitochondrial membrane to potassium; meanwhile, the second phase seems to be related to the permeability transition pore opening (Fig. 4B, f).
To directly form channels in biological membranes, the peptide BTM-P1 should cross the lipid bilayer by its N-or Cterminal sides in a potential-dependent manner according to the obtained data. We assumed that such probability is higher for the N-terminal side because of its higher hydrophobicity. To decrease the probability of a transmembrane movement of the N-terminal side of the peptide, a tryptophan residue was added to a new peptide BTM-WP1. The tryptophan indole rings have been found to be preferentially positioned near the lipid carbonyl moieties in the membrane-water interface (51,52). The interaction of synthetic tryptophan-flanked peptides with the phosphatidylcholine bilayer is energetically stronger than the hydrophobic matching, although generally it also depends on the membrane lipid composition (53). On the other hand, the flanked tryptophan residues of those peptides were not the first or the last, i.e. neither an extreme N-terminal side nor an extreme C-terminal side. We assume that the strength of the interfacial anchoring of the N-terminal tryptophan in the peptide BTM-WP1 might be even higher than due only to the preferred position of the indole ring near the lipid carbonyl moieties because the terminal amino group of the same tryptophan residue might be electrostatically attracted by a nearest glycerophosphocholine group in the membrane-water interface (54), where the dielectric constant is lower than in the water. Even a free tryptophan has been shown to dramatically affect the structural state of biomembranes and lipid bilayer of liposomes during freezing (55) that might be caused by a tandem action of its indole ring and the amino group. The results of the comparative study of the BTM-P1 and BTM-WP1 influence on the inner membrane potential and on the swelling of energized mitochondria in the KNO 3 medium are shown in Fig. 5.
No effect of BTM-WP1 on the inner membrane potential has been observed even at the concentration of 5 g/ml peptide, whereas the addition of only 2 g/ml BTM-P1 or 1 M FCCP (Fig. 5, a and b, respectively) caused complete depolarization of the membrane. BTM-WP1, at concentrations up to 6 g/ml, did not influence mitochondrial swelling in the KNO 3 medium (Fig.  5c). In contrast, the peptide BTM-P1, at a concentration of 3 g/ml, was able to cause a fast and high amplitude swelling of mitochondria in the same conditions (Fig. 5d). Both results indicate that a very simple modification of BTM-P1 can dramatically change its permeation activity, although the peptide remains as an amphipathic polycation.

DISCUSSION
Cry11Bb crystal protoxin is synthesized by B. thuringiensis subsp. medellin as a 94-kDa protein composed of 782 amino acid residues (56), and its three-dimensional structure with nine ␣ helixes in the domain I has been proposed earlier (57). This protein is toxic for mosquito larvae (44,56) after activation by mosquito larvae midgut proteases (58) and further membrane binding and permeabilization of the midgut epithelial cells (59). It has not been clear up to now which are the smallest fragments of the Cry11Bb protoxin that can permeabilize biomembranes, although Segura et al. (58) proposed that the 94-kDa protoxin is activated to form 30-and 35-kDa fragments. In the case of a 70-kDa Cry11A protoxin, for instance, the coexistence of the 32-and 36-kDa fragments has been recently reported to be essential for toxicity (60).
We assumed that a very short fragment, VAPIAKYLATA-LAKWALKQGFAKLKS (2.8 kDa), provided from the primary structure of Cry11Bb protein and containing the ␣2a helix of its domain I, might be membrane-active. This fragment has been artificially synthesized as the peptide BTM-P1, and its antimicrobial activity and a capacity to induce mitochondrial swelling have been demonstrated. 2 The data obtained in this work show that BTM-P1 uncouples respiration and oxidative phosphorylation of rat liver mitochondria. The uncoupling activity was significantly higher in the sucrose-KCl than in the sucrose medium (Fig. 1), which might be related to higher ionic strength of the sucrose-KCl medium or/and to the presence of potassium, which could penetrate the inner membrane in a voltage-dependent manner if the peptide forms potassium-permeable channels. Interestingly, at small concentrations of BTM-P1 (2 g per 1 mg of mitochondrial protein), it was still possible to observe ATP synthesis in the sucrose medium with a relatively high coefficient ADP/O and a significantly decreased respiratory control ratio (Fig. 1e).
The peptide BTM-P1 demonstrated higher uncoupling activity than some of the magainins. For comparison, a 23-amino acid residues magainin II amid, a synthetic carboxyamidated derivative of a natural polycation peptide, has been reported to completely uncouple oxidative phosphorylation at a concentration of 20 -30 g/ml and to cause nearly 50% inhibition of the FCCP-uncoupled respiration of rat liver mitochondria in sucrose medium with 5 mM succinate as substrate of oxidation (41). Magainins and their active analogues are known to form an amphiphilic ␣ helix and to permeabilize biological membranes as well as to induce channels in phospholipid membranes with slight specificity for anions (see Ref. 41 and references therein).
BTM-P1 at low concentrations dissipated the inner membrane electrical potential generated by the respiratory chain or by H ϩ -ATPase in the rat liver mitochondria. This effect of BTM-P1 is increased with an increase in concentration of KCl in isotonic incubation medium by a partial or complete substitution of sucrose with a corresponding amount of KCl. The ionic strength of the incubation medium seems to be one of the factors favoring ionophoric activity of the peptide because the high depolarization effect of BTM-P1 was also observed in isotonic choline-chloride (data not shown) and KNO 3 media (Fig. 5). Such behavior of BTM-P1 differs from that reported for some of positively charged signal sequences of mitochondrial proteins, which activated the mitochondrial multiple conductance channel and induced cyclosporin A-insensitive, potential-  (a and b) and swelling (c and d) of rat liver mitochondria. Mitochondria, 0.5 mg of protein/ml (Mc), and 2.5 M rotenone (R) were added to the incubation medium composed of 125 mM KNO 3 , 50 M EGTA, 10 mM Hepes-Tris buffer, pH 7.2, and supplemented with 5 mM succinate-Tris, pH 7.2; a and b, 10 M safranin O was added to monitor the inner membrane potential; 3WP and 5WP, 3 and 5 g/ml BTM-WP1, respectively; FCCP, 1 M FCCP; 2P and 3P, 2 and 3 g/ml BTM-P1, respectively. dependent swelling of mitochondria at concentrations of KCl lower than 50 mM; higher concentrations of KCl inhibited all these effects (40).
The dependence of the permeabilization effect of positively charged peptide BTM-P1 on the ionic strength of incubation medium might be explained taking into account the electrical field between the water bulk phase and the membrane surface, resulting from the electrical double layer of biomembranes. A surface electrical potential of approximately Ϫ60 mV was estimated for a bilayer lipid membrane containing 25-30 mol% of anionic lipids (61). According to the Guoy-Chapman theory, the thickness of the electrical double layer strongly depends on the ionic strength of the medium. The thickness of the electrical double layer in the 125 mM KCl solution is more than 3 times shorter than in 10 mM KCl (62). The average tension of the electric field near the membrane surface, within the calculated Debay-Hü ckel length ␦ 1 ϭ 0.84 nm for the KCl-medium, is 3.7 times higher than in the sucrose-medium (␦ 2 ϭ 3.09 nm), both buffered with 10 mM Hepes-Tris (Fig. 6, qualitative curves a  and b, respectively).
A higher tension of the near membrane electric field might be important for an adequate orientation of a peptide to enter into the membrane, presumably with the N-terminal side in the case of the peptide BTM-P1. For example, a simple dipole with the length of 3 nm and charges equal to 1ϩ and 1Ϫ, situated in the electric field of the near membrane space in the KCl medium, should be oriented perpendicularly to the membrane with the energy, which is 5.2 times higher than kT (k is Boltzmann constant, and T ϭ 302 K). An additional factor favoring the insertion of the polycation peptide BTM-P1 into a membrane is its net charge of 4ϩ. The peptide-motive force to insert BTM-P1 into a membrane is proportional to the net charge of the peptide and to the near membrane electric field tension.
A lower uncoupling efficiency of BTM-P1 in the sucrose medium could also be related to its electrostatic adsorption on mitochondrial membranes, as it is known for cytochrome c. Membrane adsorption of the positively charged cytochrome c decreased its electron transport activity in a low ionic strength incubation medium (49,63). Electrostatic binding of various polycations, including peptides, with anionic phospholipids of biomembranes has been suggested to induce the release of pro-apoptotic and other intermembrane proteins from mitochondria by a direct damage of the outer mitochondrial membrane or by opening of the mitochondrial permeability transition pore (64).
The peptide BTM-P1 did not induce mitochondrial swelling in the sucrose medium (Fig. 3) that was different from the effect of 3 M positively charged peptide mastoparan, which induced cyclosporin A-insensitive swelling of rat liver mitochondria in a similar incubation medium and caused potential-dependent permeabilization of liposomes (36). On the other hand, BTM-P1 induced mitochondrial swelling in high ionic strength media with the rate, amplitude, and even the kinetics of swelling depending on the type of salt used and on the metabolic state of mitochondria. Very high amplitude of swelling was observed in the KNO 3 , NH 4 NO 3 , and NaNO 3 media (Fig. 3). The effect was lower in the sucrose-KCl medium and significantly increased when sucrose was completely substituted with KCl. Relatively low amplitude and rates of swelling were observed in the Tris-NO 3 medium, indicating that the channels formed by BTM-P1 in the inner mitochondrial membrane have a restricted permeability for the cation Tris ϩ , a protonated form of Tris.
An interesting phenomenon observed in our experiments is that the rate of the BTM-P1-induced swelling was dramatically increased by energization of mitochondria incubated in the KCl or KNO 3 media, but an opposed effect was observed in the NH 4 NO 3 or Tris-NO 3 media (Fig. 3). The mechanisms of the BTM-P1-induced swelling might be different; the electrogenic transport through the formed channels for K ϩ ions but not for NH 4 ϩ ions per se. At least the channels formed by BTM-P1 seem to be less permeable for NH 4 ϩ than for potassium ions, because 1 g/ml BTM-P1 caused complete depolarization of the inner mitochondrial membrane in the KCl medium with or without inorganic phosphate (Fig. 2C) as well as in the KNO 3 medium (data not shown), and only slight transient depolarization was observed in the NH 4 NO 3 medium with inorganic phosphate (Fig.  2D, b). Because it is well known, a lipid bilayer is permeable for NH 3 ; thus, the flux of NH 4 ϩ through a membrane depends on the permeability of the membrane for H ϩ (1). The difference of pH between the mitochondrial matrix and the external medium should also affect the NH 4 ϩ flux, because a relatively high pH in the mitochondrial matrix will decrease NH 4 ϩ and increase NH 3 concentrations in it. On the other hand the steady state flux of protons from the mitochondrial matrix to the external medium, maintained by the respiratory chain functioning, might result in extrusion of NH 4 ϩ ions back to the incubation medium, thus actively preventing the high amplitude swelling of mitochondria in the NH 4 NO 3 medium. In addition, energization of mitochondria should inhibit in a potential-dependent manner the entrance of the anions NO 3 Ϫ into the mitochondrial matrix. The influence of energization on mitochondrial swelling in the Tris-NO 3 medium might be explained in the same manner, assuming that the membrane permeability for Tris ϩ also depends on the proton permeability of the inner membrane, similar to NH 4 ϩ . In the NaNO 3 medium the initial rate of swelling was somewhat higher for energized mitochondria, but further swelling was subsequently inhibited, thus resulting in lower amplitude of swelling than the observed for non-energized mitochondria (Fig. 3). This swelling inhibition might be explained if a Na ϩ /H ϩ antiport is activated during the initial phase of mitochondrial swelling, thus allowing subsequent ⌬pHdependent extrusion of sodium ions and potential-dependent extrusion of NO 3 Ϫ anions from the energized mitochondria, as a mechanism preventing the high amplitude swelling.
Mastoparan has been shown to accelerate the phosphateand Ca 2ϩ -dependent mitochondrial swelling being added to the energized mitochondria after Ca 2ϩ overloading (36). This effect might be related to the inner membrane potential drop in the presence of this peptide (36), like the caused by FCCP, allowing permeability transition pore opening (65). The influence of mastoparan addition before phosphate and Ca 2ϩ on mitochondrial swelling has not been studied. FCCP, if added together with mitochondria, prevented permeability transition pore opening by Ca 2ϩ and phosphate (Fig. 4B, c) or by BTM-P1, Ca 2ϩ , and phosphate (Fig. 4B, d) because it prevents mitochon- dria overloading with Ca 2ϩ or with Ca 2ϩ and BTM-P1. Surprisingly, BTM-P1 added before phosphate and Ca 2ϩ significantly accelerated mitochondrial permeabilization (Fig. 4), in contrast to FCCP, although uncoupling activity of BTM-P1 (Fig. 1) and the inner membrane potential drop in its presence (Fig. 2) have been demonstrated. Inorganic phosphate increased mitochondrial sensitivity to a very small concentration of BTM-P1 (0.18 M). The peptide seems to act as Ca 2ϩ in the permeability transition pore opening. After the suggestion of Pfeiffer et al. (36) for mastoparan, the polycation BTM-P1 might occupy the Ca 2ϩ binding center of the permeability transition pore, thus acting like Ca 2ϩ .
The synthetic peptide BTM-P1 with the amino acid sequence originated from a small part of the primary structure of Cry11Bb protein is composed of three fragments; a relatively hydrophobic N-terminal fragment VAPI, the fragment AKYLA-TALAKWALKQG, which has been reported as the ␣2a helix (57), and a relatively polar C-terminal fragment, FAKLKS. This peptide has predominantly random conformation in an aqueous environment and is able to adopt helicoidal conformation in a medium with higher hydrophobicity reaching 65% of helix and 35% of disordered structure in 50% trifluoroethanol. 2 In lipid bilayer membranes the polycation peptide BTM-P1 could electrostatically interact with a negatively charged surface, being situated in the hydrophobic core of the lipid monolayer, as shown in Fig. 7A, similar to the helical segment of the pore-former colicin (61).
To explain the potential-dependent permeabilization of the inner mitochondrial membrane by BTM-P1, we assume that the relatively hydrophobic N-terminal side crosses the lipid bilayer (Fig. 7B), resulting in formation of the transmembrane channel after BTM-P1 oligomerization (Fig. 7C). A high ionic strength of the incubation medium has to decrease the electrostatic interaction of the peptide lysine residues with the negatively charged surface of a membrane, thus facilitating transmembrane movement of the N-terminal side of the peptide. Energetically it seems much easier to deprotonate first the extreme amino group and to cross the membrane in a neutral form than to dehydrate a very hydrophilic amino group with a positive charge.
The positively charged oligomers formed by BTM-P1 within a membrane (Fig. 7C) has a certain probability to pass across the membrane in a potential-dependent manner before the membrane potential completely drops, thus causing direct transient permeabilization of the membrane that might explain a transient drop of the inner membrane potential after the addition of a small amount of the peptide (Fig. 2, B and D,  curves b). Subsequently, the oligomers could be reorganized, resulting in occupation of the inner surface of the inner mitochondrial membrane as well as in interaction with Ca 2ϩ binding center of the permeability transition pore. Bernardi et al. (66) have reported that Ca 2ϩ , cyclosporin A, and inhibitory Me 2ϩ ions behave as if they were competing for the same binding site(s) on the pore. It is possible that BTM-P1 also binds to the same site(s), thus activating permeability transition pore opening like Ca 2ϩ and preventing cyclosporin A binding as reported for very high concentrations of Ca 2ϩ (67).
Contrary to BTM-P1, the peptide BTM-WP1 did not affect the inner membrane potential and did not induce swelling of even energized mitochondria in the KNO 3 medium (Fig. 5). The anchoring caused by an additional N-terminal tryptophan residue, as mentioned above, seems to prevent the transmembrane movement of the N-terminal side of BTM-WP1 (Fig. 7D) and subsequent formation of the transmembrane pore.
In addition to demonstrating the biological activity of BTM-P1, our data allow speculation as to a possible mechanism of permeabilization caused by Cry protoxins at the level of the plasma membrane of midgut epithelial cells. According to the "umbrella model" (68), the pore formation by Cry family proteins results from dimerization (oligomerization) of their domains I, where only ␣4 and ␣5 helices are inserted in the membrane, and the remaining helices of the domain I are situated in the water space (umbrella). We suggest the hypothesis of "damaged umbrella oligomerization," according to which the ␣2 helix or a fragment including the ␣2 helix, like BTM-P1, could insert into the membrane maybe after ␣1 helix cleavage (47), thus allowing formation of oligomers with participation of ␣2, ␣4, and ␣5 helices. Also, it seems probable that at certain stage of proteolytic processing of the protoxin, a fragment, which includes the ␣2 helix, appears in the cytoplasm, transiently passing across the plasma membrane in oligomeric form, and reaches mitochondria to cause the mitochondrial membranes permeabilization, thus inducing apoptosis of the midgut epithelial cells.