Current models of the mode of action of Bacillus thuringiensis insecticidal crystal proteins: a critical review.

Bacillus thuringiensis (Bt) Cry toxins constitute the active ingredient in the most widely used biological insecticides and insect-resistant transgenic crops. A clear understanding of their mode of action is necessary for improving these products and ensuring their continued use. Accordingly, a long history of intensive research has established that their toxic effect is due primarily to their ability to form pores in the plasma membrane of the midgut epithelial cells of susceptible insects. In recent years, a rather elaborate model involving the sequential binding of the toxins to different membrane receptors has been developed to describe the events leading to membrane insertion and pore formation. However, it was also proposed recently that, in contradiction with this mechanism, Bt toxins function by activating certain intracellular signaling pathways which lead to the necrotic death of their target cells without the need for pore formation. Because work in this field has largely focused, for several years, on the elaboration and promotion of these two models, the present revue examines in detail the experimental evidence on which they are based. It is concluded that the presently available information still supports the notion that Bt Cry toxins act by forming pores, but most events leading to their formation, following binding of the activated toxins to their receptors, remain relatively poorly understood.

Single molecule fluorescence study of the Bacillus thuringiensis toxin Cry1Aa reveals tetramerization.

Pore-forming toxins constitute a class of potent virulence factors that attack their host membrane in a two- or three-step mechanism. After binding to the membrane, often aided by specific receptors, they form pores in the membrane. Pore formation either unfolds a cytolytic activity in itself or provides a pathway to introduce enzymes into the cells that act upon intracellular proteins. The elucidation of the pore-forming mechanism of many of these toxins represents a major research challenge. As the toxins often refold after entering the membrane, their structure in the membrane is unknown, and key questions such as the stoichiometry of individual pores and their mechanism of oligomerization remain unanswered. In this study, we used single subunit counting based on fluorescence spectroscopy to explore the oligomerization process of the Cry1Aa toxin of Bacillus thuringiensis. Purified Cry1Aa toxin molecules labeled at different positions in the pore-forming domain were inserted into supported lipid bilayers, and the photobleaching steps of single fluorophores in the fluorescence time traces were counted to determine the number of subunits of each oligomer. We found that toxin oligomerization is a highly dynamic process that occurs in the membrane and that tetramers represent the final form of the toxins in a lipid bilayer environment.

Effects of mutations within surface-exposed loops in the pore-forming domain of the Cry9Ca insecticidal toxin of Bacillus thuringiensis.

The pore-forming domain of Bacillus thuringiensis insecticidal Cry toxins is formed of seven amphipathic α-helices. Because pore formation is thought to involve conformational changes within this domain, the possible role of its interhelical loops in this crucial step was investigated with Cry9Ca double mutants, which all share the previously characterized R164A mutation, using a combination of homology modeling, bioassays and electrophysiological measurements. The mutations either introduced, neutralized or reversed an electrical charge carried by a single residue of one of the domain I loops. The ability of the 28 Cry9Ca double mutants to depolarize the apical membrane of freshly isolated Manduca sexta larval midguts was tested in the presence of either midgut juice or a cocktail of protease inhibitors because these conditions had been shown earlier to greatly enhance pore formation by Cry9Ca and its R164A single-site mutant. Most mutants retained toxicity toward neonate larvae and a pore-forming ability in the electrophysiological assay, which were comparable to those of their parental toxin. In contrast, mutants F130D, L186D and V189D were very poorly toxic and practically inactive in vitro. On the other hand, mutant E129A depolarized the midgut membrane efficiently despite a considerably reduced toxicity, and mutant Q192E displayed a reduced depolarizing ability while conserving a near wild-type toxicity. These results suggest that the conditions found in the insect midgut, including high ionic strength, contribute to minimizing the influence of surface charges on the ability of Cry9Ca and probably other B. thuringiensis toxins to form pores within their target membrane.

Midgut juice components affect pore formation by the Bacillus thuringiensis insecticidal toxin Cry9Ca.

The pore-forming ability of the Bacillus thuringiensis toxin Cry9Ca, its two single-site mutants R164A and R164K, and the 55-kDa fragment resulting from its proteolytic cleavage at R164 was evaluated under a variety of experimental conditions using an electrophysiological assay. All four toxin preparations depolarized the apical membrane of freshly isolated third-instar Manduca sexta midguts bathing in a solution containing 122 mM KCl at pH 10.5, but the 55-kDa fragment was considerably more active than Cry9Ca and its mutants. The activity of the latter toxins was greatly enhanced, however, when the experiments were conducted in the presence of fifth-instar M. sexta midgut juice. This effect was also observed after midgut juice proteins had been denatured by heating at 95 degrees C or after inorganic ions and small molecules had been removed from the midgut juice by extensive dialysis. A similar stimulation of toxin activity was also observed when the experiments were carried out in the presence of the lipids extracted from an equivalent volume of midgut juice. Depolarization of the cell membrane was also greatly enhanced, in the absence of midgut juice, by the addition of a cocktail of water-soluble protease inhibitors. These results indicate that, depending on the cleavage site and on the experimental conditions used, further proteolysis of the activated Cry9Ca toxin can either stimulate or be detrimental to its activity and that M. sexta midgut juice probably contains protease inhibitors that could play a major role in the activity of B. thuringiensis toxins in the insect midgut.

Pore-forming properties of the Bacillus thuringiensis toxin Cry9Ca in Manduca sexta brush border membrane vesicles.

The toxicity and pore-forming ability of the Bacillus thuringiensis Cry9Ca insecticidal toxin, its single-site mutants, R164A and R164K, and the 55-kDa fragment resulting from its proteolytic cleavage at residue 164 were investigated using Manduca sexta neonate larvae and fifth-instar larval midgut brush border membrane vesicles, respectively. Neither the mutations nor the proteolytic cleavage altered Cry9Ca toxicity. Compared with Cry1Ac, Cry9Ca and its mutants formed large poorly selective pores in the vesicles. Pore formation was highly dependent on pH, however, especially for wild-type Cry9Ca and both mutants. Increasing pH from 6.5 to 10.5 resulted in an irregular step-wise decrease in membrane permeabilization that was not related to a change in the ionic selectivity of the pores. Pore formation was much slower with Cry9Ca and its derivatives, including the 55-kDa fragment, than with Cry1Ac and its rate was not influenced by the presence of protease inhibitors or a reducing agent.

Mutations in domain I interhelical loops affect the rate of pore formation by the Bacillus thuringiensis Cry1Aa toxin in insect midgut brush border membrane vesicles.

Pore formation in the apical membrane of the midgut epithelial cells of susceptible insects constitutes a key step in the mode of action of Bacillus thuringiensis insecticidal toxins. In order to study the mechanism of toxin insertion into the membrane, at least one residue in each of the pore-forming-domain (domain I) interhelical loops of Cry1Aa was replaced individually by cysteine, an amino acid which is normally absent from the activated Cry1Aa toxin, using site-directed mutagenesis. The toxicity of most mutants to Manduca sexta neonate larvae was comparable to that of Cry1Aa. The ability of each of the activated mutant toxins to permeabilize M. sexta midgut brush border membrane vesicles was examined with an osmotic swelling assay. Following a 1-h preincubation, all mutants except the V150C mutant were able to form pores at pH 7.5, although the W182C mutant had a weaker activity than the other toxins. Increasing the pH to 10.5, a procedure which introduces a negative charge on the thiol group of the cysteine residues, caused a significant reduction in the pore-forming abilities of most mutants without affecting those of Cry1Aa or the I88C, T122C, Y153C, or S252C mutant. The rate of pore formation was significantly lower for the F50C, Q151C, Y153C, W182C, and S252C mutants than for Cry1Aa at pH 7.5. At the higher pH, all mutants formed pores significantly more slowly than Cry1Aa, except the I88C mutant, which formed pores significantly faster, and the T122C mutant. These results indicate that domain I interhelical loop residues play an important role in the conformational changes leading to toxin insertion and pore formation.

Helix alpha 4 of the Bacillus thuringiensis Cry1Aa toxin plays a critical role in the postbinding steps of pore formation.

Helix alpha 4 of Bacillus thuringiensis Cry toxins is thought to play a critical role in the toxins' mode of action. Accordingly, single-site substitutions of many Cry1Aa helix alpha 4 amino acid residues have previously been shown to cause substantial reductions in the protein's pore-forming activity. Changes in protein structure and formation of intermolecular disulfide bonds were investigated as possible factors responsible for the inactivity of these mutants. Incubation of each mutant with trypsin and chymotrypsin for 12 h did not reveal overt structural differences with Cry1Aa, although circular dichroism was slightly decreased in the 190- to 210-nm region for the I132C, S139C, and V150C mutants. The addition of dithiothreitol stimulated pore formation by the E128C, I132C, S139C, T142C, I145C, P146C, and V150C mutants. However, in the presence of these mutants, the membrane permeability never reached that measured for Cry1Aa, indicating that the formation of disulfide bridges could only partially explain their loss of activity. The ability of a number of inactive mutants to compete with wild-type Cry1Aa for pore formation in brush border membrane vesicles isolated from Manduca sexta was also investigated with an osmotic swelling assay. With the exception of the L147C mutant, all mutants tested could inhibit the formation of pores by Cry1Aa, indicating that they retained receptor binding ability. These results strongly suggest that helix alpha 4 is involved mainly in the postbinding steps of pore formation.

Chemical modification of Bacillus thuringiensis Cry1Aa toxin single-cysteine mutants reveals the importance of domain I structural elements in the mechanism of pore formation.

Bacillus thuringiensis Cry toxins form pores in the apical membrane of insect larval midgut cells. To investigate their mechanism of membrane insertion, mutants in which cysteine replaced individual amino acids located within the pore-forming domain of Cry1Aa were chemically modified with sulfhydryl-specific reagents. The thiol group of cysteine was highly susceptible to oxidation and its reactivity was significantly increased when the toxins were purified under reducing conditions. Addition of a biotin group to the cysteine had little effect on the ability of the toxins to permeabilize Manduca sexta brush border membrane vesicles except for a slight reduction in activity for S252C and a large increase in activity for Y153C. The activity of Y153C was also significantly increased after modification by reagents that added an aromatic or a charged group to the cysteine. When permeability assays were performed in the presence of streptavidin, a large biotin-binding protein, the pore-forming activity of several mutants, including Y153C, where the altered residue is located within the hairpin comprising helices alpha4 and alpha5, or in adjacent loops, was significantly reduced. These results support the umbrella model of toxin insertion.

Atomic force microscopy imaging of Bacillus thuringiensis Cry1 toxins interacting with insect midgut apical membranes.

Atomic force microscopy was used to image Bacillus thuringiensis (Bt) toxins interacting with their natural targets, Manduca sexta midgut brush border membranes (BBMs), as well as with dipalmitoylphosphatidylcholine-dioleoylphosphatidylcholine (DPPC-DOPC) solid-supported lipid bilayers. In lipid bilayers, Cry1Aa formed structures 30-60 nm wide and 3-7 nm high, mostly at the interface of domains formed by the two different lipids or at the edge of DOPC-enriched domains. BBM vesicles, in the absence of toxin, formed flat membrane fragments of up to 25 microm(2) and 4.2 nm high, with irregular embedded structures. After incubation with Cry1Aa, Cry1Ac and Cry1C, which are active against M. sexta, new structures, 35 nm wide and 5.1-6.7 nm high, were observed in some membrane fragments, sometimes only in particular regions. Their density, which reached a plateau within 4 h, was toxin- and concentration-dependent. The structures formed by Cry1Ac were often grouped into dense, two-dimensional arrangements. No such specific interactions were observed with Cry1Ba, which is inactive against M. sexta. This study provides the first visual demonstration of specific interactions of Bt toxins with insect midgut BBMs at the nanometric scale. The observed structures likely represent the protein complexes forming functional Bt pores in target membranes.

Cysteine scanning mutagenesis of alpha4, a putative pore-lining helix of the Bacillus thuringiensis insecticidal toxin Cry1Aa.

Helix alpha4 of Bacillus thuringiensis Cry toxins is thought to line the lumen of the pores they form in the midgut epithelial cells of susceptible insect larvae. To define its functional role in pore formation, most of the alpha4 amino acid residues were replaced individually by a cysteine in the Cry1Aa toxin. The toxicities and pore-forming abilities of the mutated toxins were examined, respectively, by bioassays using neonate Manduca sexta larvae and by a light-scattering assay using midgut brush border membrane vesicles isolated from M. sexta. A majority of these mutants had considerably reduced toxicities and pore-forming abilities. Most mutations causing substantial or complete loss of activity map on the hydrophilic face of the helix, while most of those having little or only relatively minor effects map on its hydrophobic face. The properties of the pores formed by mutants that retain significant activity appear similar to those of the pores formed by the wild-type toxin, suggesting that mutations resulting in a loss of activity interfere mainly with pore formation.

Role of NHERF1, cystic fibrosis transmembrane conductance regulator, and cAMP in the regulation of aquaporin 9.

Water and solute transport across the plasma membrane of cells is a crucial biological function that is mediated mainly by aquaporins and aquaglyceroporins. The regulation of these membrane proteins is still incompletely understood. Using the male reproductive tract as a model system in which water and glycerol transport are critical for the establishment of fertility, we now report a novel pathway for the regulation of aquaporin 9 (AQP9) permeability. AQP9 is the major aquaglyceroporin of the epididymis, liver, and peripheral leukocytes, and its COOH-terminal portion contains a putative PDZ binding motif (SVIM). Here we show that NHERF1, cystic fibrosis transmembrane conductance regulator (CFTR), and AQP9 co-localize in the apical membrane of principal cells of the epididymis and the vas deferens, and that both NHERF1 and CFTR co-immunoprecipitate with AQP9. Overlay assays revealed that AQP9 binds to both the PDZ1 and PDZ2 domains of NHERF1, with an apparently higher affinity for PDZ1 versus PDZ2. Pull-down assays showed that the AQP9 COOH-terminal SVIM motif is essential for interaction with NHERF1. Functional assays on isolated tubules perfused in vitro showed a high permeability of the apical membrane to glycerol, which is inhibited by the AQP9 inhibitor, phloretin, and is markedly activated by cAMP. The CFTR inhibitors DPC, GlyH-101 and CFTRinh-172 all significantly reduced the cAMP-activated glycerol-induced cell swelling. We propose that CFTR is an important regulator of AQP9 and that the interaction between AQP9, NHERF1, and CFTR may facilitate the activation of AQP9 by cAMP.

Effect of insect larval midgut proteases on the activity of Bacillus thuringiensis Cry toxins.

To test the possibility that proteolytic cleavage by midgut juice enzymes could enhance or inhibit the activity of Bacillus thuringiensis insecticidal toxins, once activated, the effects of different toxins on the membrane potential of the epithelial cells of isolated Manduca sexta midguts in the presence and absence of midgut juice were measured. While midgut juice had little effect on the activity of Cry1Aa, Cry1Ac, Cry1Ca, Cry1Ea, and R233A, a mutant of Cry1Aa from which one of the four salt bridges linking domains I and II of the toxin was eliminated, it greatly increased the activity of Cry1Ab. In addition, when tested in the presence of a cocktail of protease inhibitors or when boiled, midgut juice retained almost completely its capacity to enhance Cry1Ab activity, suggesting that proteases were not responsible for the stimulation. On the other hand, in the absence of midgut juice, the cocktail of protease inhibitors also enhanced the activity of Cry1Ab, suggesting that proteolytic cleavage by membrane proteases could render the toxin less effective. The lower toxicity of R233A, despite a similar in vitro pore-forming ability, compared with Cry1Aa, cannot be accounted for by an increased susceptibility to midgut proteases. Although these assays were performed under conditions approaching those found in the larval midgut, the depolarizing activities of the toxins correlated only partially with their toxicities.

Kinetics of pore formation by the Bacillus thuringiensis toxin Cry1Ac.

After binding to specific receptors, Cry toxins form pores in the midgut apical membrane of susceptible insects. The receptors could form part of the pore structure or simply catalyze pore formation and consequently be recycled. To discriminate between these possibilities, the kinetics of pore formation in brush border membrane vesicles isolated from Manduca sexta was studied with an osmotic swelling assay. Pore formation, as deduced from changes in membrane permeability induced by Cry1Ac during a 60-min incubation period, was strongly dose-dependent, but rapidly reached a maximum as toxin concentration was increased. Following exposure of the vesicles to the toxin, the osmotic swelling rate reached a maximum shortly after a delay period. Under these conditions, at relatively high toxin concentrations, the maximal osmotic swelling rate increased linearly with toxin concentration. When vesicles were incubated for a short time with the toxin and then rapidly cooled to prevent the formation of new pores before and during the osmotic swelling experiment, a plateau in the rate of pore formation was observed as toxin concentration was increased. Taken together, these results suggest that the receptors do not act as simple catalysts of pore formation, but remain associated with the pores once they are formed.

A mechanical force contributes to the "osmotic swelling" of brush-border membrane vesicles.

Brush-border membrane vesicles and an osmotic swelling assay have been used extensively to monitor the pore-forming activity of Bacillus thuringiensis toxins. After a hypertonic shock, Manduca sexta midgut brush-border membrane vesicles shrink rapidly and reswell partially to a volume that depends on membrane permeability and toxin concentration rather than regaining their original volume as expected from theoretical models. Because efflux of buffer from the vesicles, as they shrink, could contribute to this phenomenon, vesicles were mixed with a hypertonic solution of the buffer with which they were loaded. Under these conditions, they are not expected to reswell, since the same solute is present on both sides of the membrane. Nevertheless, with several buffers, vesicles reswelled readily, an observation that demonstrates the involvement of an additional restoration force. Reswelling also occurred when, in the absence of toxin, the buffers were replaced by glucose, a solute that diffuses readily across the membrane, but did not occur with rat liver microsomes, despite their permeability to glucose. Unexpected swelling was also observed with rabbit jejunum brush-border membrane vesicles, suggesting that the cytoskeleton, present in brush-border membrane vesicles but absent from microsomes, could be responsible for the restoration force.

Influence of the biophysical and biochemical environment on the kinetics of pore formation by Cry toxins.

The effect of Bacillus thuringiensis toxins on the permeability of the luminal membrane of Manduca sexta midgut columnar epithelial cells is strongly influenced by several biophysical and biochemical factors, including pH, ionic strength, and divalent cations, suggesting an important role for electrostatic interactions. The influence of these factors can differ greatly, however, depending on the toxin being studied, even for closely related toxins such as Cry1Ac and Cry1Ca. In the present study, the possibility of using temperature changes as a tool for controlling the rate and extent of pore formation in midgut brush border membrane vesicles was evaluated. Lowering temperature gradually decreased the rate of pore formation, but had little effect on the permeability of vesicles previously incubated with toxin at room temperature. The formation of new pores, following incubation of the vesicles with toxin, could thus be almost abolished by rapidly cooling the vesicles to 2 degrees C. Using this approach, changes in the rate of pore formation could be more easily distinguished from alterations in the properties of the pores formed, thus allowing a more detailed analysis of the kinetics and mechanism of pore formation.

Protease inhibitors fail to prevent pore formation by the activated Bacillus thuringiensis toxin Cry1Aa in insect brush border membrane vesicles.

To investigate whether membrane proteases are involved in the activity of Bacillus thuringiensis insecticidal toxins, the rate of pore formation by trypsin-activated Cry1Aa was monitored in the presence of a variety of protease inhibitors with Manduca sexta midgut brush border membrane vesicles and by a light-scattering assay. Most of the inhibitors tested had no effect on the pore-forming ability of the toxin. However, phenylmethylsulfonyl fluoride, a serine protease inhibitor, promoted pore formation, although this stimulation only occurred at higher inhibitor concentrations than those commonly used to inhibit proteases. Among the metalloprotease inhibitors, o-phenanthroline had no significant effect; EDTA and EGTA reduced the rate of pore formation at pH 10.5, but only EDTA was inhibitory at pH 7.5. Neither chelator affected the properties of the pores already formed after incubation of the vesicles with the toxin. Taken together, these results indicate that, once activated, Cry1Aa is completely functional and does not require further proteolysis. The effect of EDTA and EGTA is probably better explained by their ability to chelate divalent cations that could be necessary for the stability of the toxin's receptors or involved elsewhere in the mechanism of pore formation.

Helix 4 mutants of the Bacillus thuringiensis insecticidal toxin Cry1Aa display altered pore-forming abilities.

The role played by alpha-helix 4 of the Bacillus thuringiensis toxin Cry1Aa in pore formation was investigated by individually replacing each of its charged residues with either a neutral or an oppositely charged amino acid by using site-directed mutagenesis. The majority of the resulting mutant proteins were considerably less toxic to Manduca sexta larvae than Cry1Aa. Most mutants also had a considerably reduced ability to form pores in midgut brush border membrane vesicles isolated from this insect, with the notable exception of those with alterations at amino acid position 127 (R127N and R127E), located near the N-terminal end of the helix. Introducing a negatively charged amino acid near the C-terminal end of the helix (T142D and T143D), a region normally devoid of charged residues, completely abolished pore formation. For each mutant that retained detectable pore-forming activity, reduced membrane permeability to KCl was accompanied by an approximately equivalent reduction in permeability to N-methyl-D-glucamine hydrochloride, potassium gluconate, sucrose, and raffinose and by a reduced rate of pore formation. These results indicate that the main effect of the mutations was to decrease the toxin's ability to form pores. They provide further evidence that alpha-helix 4 plays a crucial role in the mechanism of pore formation.

Patch-clamp study of the apical membrane of the midgut of Manduca sexta larvae: direct demonstration of endogenous channels and effect of a Bacillus thuringiensis toxin.

The patch-clamp technique was applied to the apical membrane of epithelial midgut cells of a lepidoptera, Manduca sexta L. Access to the apical membrane, the main target site of Bacillus thuringiensis (Bt) toxins, was achieved by using freshly isolated larval midgut preparations mounted onto holding glass pipettes. The epithelial cells retained their functional integrity, as evidenced by the magnitude of intracellular potentials recorded with microelectrodes. With standard 32 mM K(+) solution in the bath and the patch-clamp pipette, endogenous channel activity was detected in about 50% of experiments, mainly in moulting larvae and larvae that had been kept at reduced temperature for at least two days prior to the experiments. In both cell-attached and inside-out patch-clamp configurations, different types of channel were observed, with conductances varying between about 5 and 50 pS and different conducting properties. Addition of trypsin-activated Cry1Ac Bt toxin in the patch-clamp pipette triggered, after a delay, large conductances of a few nanosiemens. This is the first study allowing exploration, in the intact midgut, of the properties of apical membrane channels and the direct interaction between the apical membrane of epithelial cells and pathogenic agents such as Bt toxins.

Estimation of the radius of the pores formed by the Bacillus thuringiensis Cry1C delta-endotoxin in planar lipid bilayers.

Pore formation constitutes a key step in the mode of action of Bacillus thuringiensis delta-endotoxins and various activated Cry toxins have been shown to form ionic channels in receptor-free planar lipid bilayers at high concentrations. Multiple conductance levels have been observed with several toxins, suggesting that the channels result from the multimeric assembly of a variable number of toxin molecules. To test this possibility, the size of the channels formed by Cry1C was estimated with the non-electrolyte exclusion technique and polyethylene glycols of various molecular weights. In symmetrical 300 mM KCl solutions, Cry1C induced channel activity with 15 distinct conductance levels ranging from 21 to 246 pS and distributed in two main conductance populations. Both the smallest and largest conductance levels and the mean conductance values of both populations were systematically reduced in the presence of polyethylene glycols with hydrated radii of up to 1.05 nm, indicating that these solutes can penetrate the pores formed by the toxin. Larger polyethylene glycols had little effect on the conductance levels, indicating that they were excluded from the pores. Our results indicate that Cry1C forms clusters composed of a variable number of channels having a similar pore radius of between 1.0 and 1.3 nm and gating synchronously.

Role of helix 3 in pore formation by the Bacillus thuringiensis insecticidal toxin Cry1Aa.

Helix 3 of the Cry1Aa toxin from Bacillus thuringiensis possesses eight charged amino acids. These residues, with the exception of those involved in intramolecular salt bridges (E90, R93, E112, and R115), were mutated individually either to a neutral or to an oppositely charged amino acid. The mutated genes were expressed, and the resultant, trypsin-activated toxins were assessed for their toxicity to Manduca sexta larvae and their ability to permeabilize M. sexta larval midgut brush border membrane vesicles to KCl, sucrose, raffinose, potassium gluconate, and N-methyl-D-glucamine hydrochloride with a light-scattering assay based on osmotic swelling. Most mutants were considerably less toxic than Cry1Aa. Replacing either E101, E116, E118, or D120 by cysteine, glutamine, or lysine residues had only minor effects on the properties of the pores formed by the modified toxins. However, half of these mutants (E101C, E101Q, E101K, E116K, E118C, and D120K) had a significantly slower rate of pore formation than Cry1Aa. Mutations at R99 (R99C, R99E, and R99Y) resulted in an almost complete loss of pore-forming ability. These results are consistent with a model in which alpha-helix 3 plays an important role in the mechanism of pore formation without being directly involved in determining the properties of the pores.