Manduca sexta midgut brush border membrane vesicles proceeds by more

Transcription

1 Journal of Cell Science 11, (1997) Printed in Great Britain The Company of Biologists Limited 1997 JCS The Bacillus thuringiensis Cry1Ac toxin-induced permeability change in Manduca sexta midgut brush border membrane vesicles proceeds by more than one mechanism Joe Carroll 1, Michael G. Wolfersberger 2 and David J. Ellar 1, * 1 Department of Biochemistry, University of Cambridge, Cambridge CB2 1GA, UK 2 Department of Biology, Temple University, Philadelphia, Pennsylvania 19122, USA *Author for correspondence ( dje1@mole.bio.cam.ac.uk) SUMMARY Aminopeptidase N purified from whole Manduca sexta midgut binds the Cry1Ac insecticidal toxin from Bacillus thuringiensis and this binding is inhibited by N-acetylgalactosamine (GalNAc). We have examined the membrane permeabilising activity of the Cry1Ac toxin using brush border membrane vesicles (BBMV) prepared from the anterior (A-BBMV) and posterior (P-BBMV) subregions of the M. sexta midgut. A toxin mixing assay demonstrated a faster rate of toxin activity on P-BBMV than on A-BBMV. In the presence of GalNAc this rapid activity on P-BBMV was reduced to the rate seen with A-BBMV. GalNAc had no effect on the rate of A-BBMV permeabilisation by Cry1Ac. Aminopeptidase N assays of A- and P-BBMV demonstrated that this Cry1Ac binding protein is concentrated in the posterior midgut region of M. sexta. It therefore appears that there are two mechanisms by which Cry1Ac permeabilises the M. sexta midgut membrane: a GalNAc-sensitive mechanism restricted to the posterior midgut region, probably involving aminopeptidase N binding, and a previously undetected mechanism found in both the posterior and anterior regions. Key words: Bacillus thuringiensis, Cry1Ac, Brush border membrane vesicle, Membrane, Permeabilisation, Manduca sexta, Aminopeptidase N, N-acetylgalactosamine INTRODUCTION Bacillus thuringiensis synthesises a family of δ-endotoxin proteins (Cry toxins: named according to the revised Cry holotype toxin nomenclature; WWW site: that form insoluble inclusions during sporulation and exhibit toxicity to a range of invertebrates (Feitelson et al., 1992). Extensive studies with insects have shown that after ingestion the Cry toxin inclusion is solubilised and activated in the midgut, ultimately producing a toxin-induced lysis of susceptible midgut epithelial cells (Knowles, 1994). It was proposed that cytolysis proceeds via the initial interaction of an activated toxin with specific membrane receptors, followed by the formation of a 1-2 nm diameter pore leading to cell death by colloid-osmotic lysis (Knowles and Ellar, 1987). This premise was supported by experiments using brush border membrane vesicles (BBMV) prepared from insect midguts, which demonstrated the presence of high affinity receptors for Cry toxins in susceptible insects (Hofmann et al., 1988; Van Rie et al., 1989). The characteristics of the toxin lesion formed in the membrane after the initial interaction are still disputed and may vary depending on the toxin:insect system under investigation (Knowles and Dow, 1993). Using Manduca sexta midgut BBMV, the Cry1Ac toxin was observed to alter the membrane permeability for monovalent cations, anions and neutral solutes (Carroll and Ellar, 1993). Recent evidence suggests that Cry1Ac can form a 2 nm diameter pore in M. sexta midgut BBMV (Carroll and Ellar, 1997) or in a planar lipid bilayer to which these BBMV were fused (Martin and Wolfersberger, 1995). While specific in vivo, Cry toxins have been shown to form channels in lipid bilayers in the absence of putative receptor molecules (Grochulski et al., 1995; Schwartz et al., 1993; Slatin et al., 199). However, fusing receptor-bearing BBMV to these lipid bilayers decreases greatly the toxin concentration needed to produce membrane conductance changes (Lorence et al., 1995; Martin and Wolfersberger, 1995). In addition, the method of in vitro toxin activation can influence the channel forming activity of Cry toxins in lipid bilayers (Smedley et al., 1997). Putative Cry toxin receptors have been detected in insect midgut membranes (Belfiore et al., 1994; Knowles et al., 1991; Martínez-Ramírez et al., 1994; Oddou et al., 1991; Sanchis and Ellar, 1993; Vadlamudi et al., 1993). In four lepidopteran insects, M. sexta (Knight et al., 1994; Sangadala et al., 1994), Heliothis virescens (Gill et al., 1995), Lymantria dispar (Valaitis et al., 1995) and Plutella xylostella (Luo et al., 1997), aminopeptidase N has been identified as a Cry1Ac receptor. Evidence suggests that GalNAc (N-acetylgalactosamine)

2 31 J. Carroll, M. G. Wolfersberger and D. J. Ellar residues on aminopeptidase N play a role in the interaction with the Cry1Ac toxin (Gill et al., 1995; Knight et al., 1994; Masson et al., 1995). Curiously, while aminopeptidase N activity is localised primarily in the posterior region of the M. sexta larval midgut membrane (Wolfersberger, 1996), immunocytochemical analysis suggested that Cry1Ac binds uniformly along the entire length of the midgut (Bravo et al., 1992a). It was therefore of interest to determine whether the pore forming activity of Cry1Ac follows the distribution of aminopeptidase N or exhibits some other pattern of activity along the M. sexta midgut. MATERIALS AND METHODS Cry1Ac purification and activation Bacillus thuringiensis subsp. kurstaki HD73, which synthesises a single Cry1Ac protoxin inclusion (Höfte and Whiteley, 1989) was obtained from USDA Northern Regional Research Laboratories, Illinois, USA. Growth conditions were as described for Bacillus megaterium KM (Stewart et al., 1981). The Cry1Ac crystal inclusion was purified from sporulated B. thuringiensis cultures using discontinuous sucrose gradients (Thomas and Ellar, 1983) and the protein concentration was determined by the method of Lowry et al. (1951) using bovine serum albumin (BSA) as a standard. Solubilisation, activation and sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) analysis of the Cry1Ac toxin was as described previously (Carroll and Ellar, 1993). Quantification of the activated toxin was performed by gel scanning densitometry of SDS- PAGE-separated and Coomassie blue-stained protein (Carroll and Ellar, 1997). Manduca sexta midgut BBMV preparation Manduca sexta eggs were obtained from Dr S. Reynolds (University of Bath, School of Biological Sciences, United Kingdom) and the larvae reared using an artificial diet (Bell and Joachim, 1976). BBMV were isolated from the anterior (A-BBMV) and posterior (P-BBMV) midgut regions using a slightly modified method of Wolfersberger et al. (1987) as described by Carroll and Ellar (1993). The central midgut region was discarded. BBMV protein concentration was quantified by the method of Bradford (1976) using a Bio-Rad protein assay dye reagent. Aminopeptidase N activity was assayed at 25 C according to the method of Hafkensheid (1984) using the chromogenic substrate L- leucine-p-nitroanilide (Sigma). The absorbance change at 45 nm was followed and p-nitroanilide (Sigma) standard concentrations were used to calculate specific enzyme activities. The millimolar absorption coefficient of p-nitroanilide was taken to be 9.9 litre mmol 1 cm 1. BBMV permeability assay BBMV solute permeability was studied using a light scattering assay described by Carroll and Ellar (1993) with minor modifications. BBMV were used at a final concentration of.2 mg/ml and the buffer was 1 mm 2-[cyclohexylamino]-ethanesulphonic acid (CHES)/KOH/.1% (w/v) BSA, ph 9. The hyperosmotic solute used was 15 mm KCl in 1 mm CHES/KOH/.1% (w/v) BSA, ph 9, which was mixed 1:1 (v/v) with the BBMV suspension to give a final KCl concentration of 75 mm. Both toxin pre-incubation and toxin mixing assays (Carroll and Ellar, 1993) were used. The pre-incubation experiments employed a 1 hour toxin-bbmv incubation period prior to exposing the treated vesicles to hyperosmotic KCl. By contrast, the mixing assay monitored changes from the point where toxin in a hyperosmotic KCl solution was mixed with BBMV. The action of the sugars GalNAc and GlcNAc (N-acetylglucosamine) on toxin activity in the mixing assay was assessed at a final experimental concentration of 1 mm. Non-toxin-mediated changes in BBMV exposed to hyperosmotic solute solutions were subtracted from toxin effects. Toxin-induced signal recoveries were then determined from the difference data and are shown relative to the change observed for the control. Comparisons of linear regression lines were performed by an analysis of covariance using the computer program COMPREG (Wiggans et al., 1983). A P value of less than.5 was considered statistically significant. SDS-PAGE and immunoblotting SDS-PAGE separation of BBMV proteins was performed by a modified method of Laemmli and Favre (1973) as described by Thomas and Ellar (1983). Electro-transfer of proteins from gels to nitrocellulose paper was carried out with a semi-dry blot apparatus (Department of Biochemistry, University of Cambridge, UK) using a blot buffer of 39 mm glycine, 48 mm Tris,.375% (w/v) SDS, 1% (v/v) methanol. Cry1Ac toxin binding proteins were then identified by the method of Knowles et al. (1991). Proteolytically activated Cry1Ac was incubated with the nitrocellulose paper at 2 µg/ml for 6 minutes. Bound toxin was detected using a primary (rabbit anti-toxin) antibody raised to an activated Cry1Ac toxin preparation, followed by a secondary (peroxidase-conjugated goat anti-rabbit) antibody (Sigma). The peroxidase colour reaction was developed as described by Hawkes et al. (1982). All incubations were carried out in Trisbuffered saline (15 mm NaCl, 1 mm Tris-HCl, ph 7.4) containing 3% (w/v) bovine haemoglobin, the nitrocellulose being washed with Tris-buffered saline between incubations. RESULTS Aminopeptidase N activity Aminopeptidase N activity measured for A-BBMV was only 2.5% of that found in P-BBMV when expressed per mg BBMV protein. The specific activities (±s.d.) were.35 ±.23 µmol minute 1 mg 1 (n=4) for A-BBMV and 14.18±2.84 µmol minute 1 mg 1 (n=5) for P-BBMV. This compares favourably with an earlier report using larvae from a different colony (Wolfersberger, 1996). Cry1Ac-induced BBMV permeability changes Toxin-induced changes in BBMV permeability were measured using a light-scattering assay (Carroll and Ellar, 1993). A recovery in the light-scattering signal after vesicle shrinkage in a hyperosmotic solute occurs as BBMV re-swell due to entry of the solute and water into the vesicles. Toxin data were corrected for signal recoveries observed in control BBMV. Cry1Ac pre-incubation induced an increased permeability for KCl in M. sexta A- and P-BBMV (Fig. 1). While data for most toxin concentrations tested show a greater action of Cry1Ac on A-BBMV, the regression analysis performed on the accumulated data does not give a significant difference between the permeability changes induced by Cry1Ac on A- and P-BBMV (Fig. 1 inset). Pre-incubation experiments analyse a toxin lesion preformed in the BBMV before mixing with hyperosmotic solution. In contrast, toxin mixing experiments monitor changes from the point of introducing toxin to a BBMV suspension, and therefore follow the time course of toxin-induced permeability changes including the initial toxin-receptor interaction. Initial experiments using this approach suggested that Cry1Ac acted faster on P-BBMV (Fig. 2). GalNAc and GlcNAc effects on Cry1Ac activity The action of the similar sugars GalNAc and GlcNAc on toxininduced permeability changes in M. sexta BBMV was

3 Toxin-induced membrane permeabilisation A-BBMV P-BBMV Cry1Ac (pmol/mg BBMV) Time (seconds) Fig. 1. Cry1Ac-induced signal recovery in Manduca sexta BBMV swelling after mixing with hyperosmotic KCl: pre-incubation experiment. BBMV (.2 mg/ml) equilibrated with 1 mm CHES/KOH/.1% (w/v) BSA, ph 9., were incubated with Cry1Ac toxin (36 pmol/mg BBMV) for 6 minutes at 2-21 C. The effect of Cry1Ac on BBMV re-swelling was observed after mixing an equal volume of treated BBMV with 1 mm CHES/KOH/.1% (w/v) BSA containing 15 mm KCl, ph 9., at 2-21 C using a stopped-flow spectrometer and followed using 9 light scattering at 45 nm. The traces are the average of three separate determinations corrected for control changes, showing light-scattering signal recoveries relative to the signal change observed for the control. Inset: linear regression analysis of Cry1Ac concentration against BBMV re-swelling at 1 seconds. Filled circles, A-BBMV; open squares, P-BBMV. Data is for signal recoveries between 1% and 8%; Cry1Ac concentrations giving zero or saturating recoveries were not used. The lines for A- BBMV (solid line, y=47x+2.14) and P-BBMV (broken line, y=47.2x 9.53) are not significantly different (P>.5 by analysis of covariance). Each point represents data obtained from the average of three separate determinations corrected for control changes, the lightscattering signal recoveries expressed relative to the signal change observed for the control P-BBMV A-BBMV Time (seconds) Fig. 2. Cry1Ac-induced signal recovery in Manduca sexta BBMV swelling after mixing with hyperosmotic KCl: toxin mixing assay. BBMV (.2 mg/ml) equilibrated with 1 mm CHES/KOH/.1% (w/v) BSA, ph 9., were mixed with 1 mm CHES/KOH/.1% (w/v) BSA plus 15 mm KCl, ph 9., containing Cry1Ac toxin (44 pmol/mg BBMV) using a stopped-flow spectrometer at 2-21 C. Reswelling was followed as the change in 9 light scattering at 45 nm over 2 seconds. Each trace represents the average of two separate determinations corrected for control changes, showing lightscattering signal recoveries relative to the signal change observed for the control. Cry1Ac toxin bound to a 12 kda protein in P-BBMV. Binding of Cry1Ac to a protein of this molecular mass was not observed for A-BBMV. Only after prolonging the final signal development period of this blotting procedure was a possible binding component observed in A-BBMV (data not shown). The molecular mass of this band was not determined, as it either only just entered the separating gel or was at the interface between the stacking and separating gels. This Cry1Ac binding component in A-BBMV was not seen in P-BBMV. examined. Fig. 3 demonstrates that GalNAc inhibited the action of Cry1Ac on P-BBMV relative to GlcNAc. In contrast, no significant difference was seen between Cry1Ac-induced swelling traces for A-BBMV in the presence of either GalNAc or GlcNAc (Fig. 3). The Cry1Ac activity on P-BBMV in the presence of GalNAc appeared to follow a similar time course to toxin-induced permeability changes on A-BBMV. The relationship between toxin concentration and the time taken for a 1% recovery in the toxin-induced BBMV swelling signal [T 1% (seconds)] supports this observation (Fig. 4). The variable [Cry1Ac] 1 expressed as pmol toxin per mg BBMV, was used to linearise the data, making the comparison simpler. Both sugars slowed the Cry1Ac-induced swelling recovery when compared with the data obtained for KCl alone (data not shown). However, the effect was not specific for either sugar and the comparison between swelling changes occurring with hyperosmotic KCl in the presence or absence of sugar is complicated by the differences in osmotic strength under these conditions. Cry1Ac binding proteins The presence of toxin binding proteins in either A-BBMV or P-BBMV was investigated. Fig. 5 demonstrates that the DISCUSSION Aminopeptidase N, a Cry1Ac toxin binding protein in the midgut of M. sexta larvae (Knight et al., 1994; Sangadala et al., 1994), was observed to be highly concentrated in the posterior midgut region of this insect (Wolfersberger, 1996; this report). The presumption that Cry1Ac would therefore exert a greater effect on BBMV from this midgut area was not supported by experiments looking at permeability changes in BBMV from anterior and posterior midgut regions after toxin pre-incubation (Fig. 1). While care should be exercised when interpreting changes in BBMV light scattering signals and the effect of a toxin pre-incubation step on different BBMV populations (Carroll and Ellar, 1997; Meissner, 1988), Cry1Ac induced a marked change in both A- and P-BBMV permeability for KCl (Fig. 1). Pre-incubation experiments are designed to look at the final toxin lesion formed during the prolonged incubation period. The possible recycling of the receptor-toxin interaction or the involvement of different interactions proceeding at different rates may not be obvious using this type of experiment. In contrast, toxin mixing experiments follow the time course of permeability change, which involves the initial receptor inter-

4 312 J. Carroll, M. G. Wolfersberger and D. J. Ellar A B +GalNAc +GlcNAc +GlcNAc +GalNAc Time (seconds) Fig. 3. Effect of GalNAc and GlcNAc on Cry1Ac-induced signal recoveries in Manduca sexta BBMV swelling after mixing with hyperosmotic KCl: toxin mixing assay. BBMV (.2 mg/ml) equilibrated with 1 mm CHES/KOH/.1% (w/v) BSA, ph 9., were mixed with 1 mm CHES/KOH/.1% (w/v) BSA plus 15 mm KCl and 2 mm GalNAc or GlcNAc, ph 9., containing Cry1Ac toxin (39.4 pmol/mg BBMV), using a stopped-flow spectrometer at 2-21 C. Re-swelling was followed as the change in 9 light scattering at 45 nm over 2 seconds. Each trace represents the average of two separate determinations corrected for control changes, showing light-scattering signal recoveries relative to the signal change observed for the control. (A) P-BBMV; (B) A-BBMV. action, toxin-membrane intercalation and subsequent membrane permeability changes (Carroll and Ellar, 1993). Using this approach, Cry1Ac appeared to exert a permeabilising activity on P-BBMV at a faster rate than on A-BBMV (Fig. 2). A simple comparison of data showing the action of Cry1Ac in the presence of GalNAc and GlcNAc was performed (Fig. 4). Using GalNAc, a sugar that inhibits the Cry1Ac interaction with aminopeptidase N (Knight et al., 1994; Masson et al., 1995), the toxin permeabilising activity on P-BBMV in a mixing assay was significantly reduced when compared to changes in the presence of GlcNAc (Figs 3, 4). However, no difference was observed between the Cry1Ac-induced permeability changes for A-BBMV in the presence of either sugar (Figs 3, 4). Because the residual P-BBMV swelling induced by Cry1Ac in the presence of GalNAc was similar to that seen for Cry1Ac on A-BBMV when either sugar was present, it appears that there are at least two mechanisms by which Cry1Acinduced pore formation proceeds in the M. sexta midgut. One mechanism is GalNAc inhibitable and seems to be highly concentrated in the posterior region of the midgut. The other is GalNAc insensitive and appears to be common to both anterior and posterior midgut regions. Cry1Ac bound to a 12 kda protein in P-BBMV but not in A-BBMV (Fig. 5). M. sexta aminopeptidase N migrates as a protein of 12 kda when subjected to SDS/PAGE (Knight et T 1% (seconds) [Cry1Ac] -1 Fig. 4. Effect of GalNAc and GlcNAc on the time taken for a Cry1Ac-induced 1% signal recovery in Manduca sexta BBMV swelling after mixing with hyperosmotic KCl: toxin mixing assay. Conditions as described in legend for Fig. 3. Cry1Ac concentrations are in units of pmol toxin per mg BBMV. Circles, A-BBMV; squares, P-BBMV. Open symbols represent GalNAc data and filled symbols GlcNAc data. Linear regression analysis was performed: A-BBMV- GlcNAc, solid line; A-BBMV-GalNAc, dotted line; P-BBMV- GlcNAc, dot and dashed line; P-BBMV-GalNAc, dashed line. The regression lines for A-BBMV-GlcNAc, A-BBMV-GalNAc and P- BBMV-GalNAc are not significantly different (P>.5 by analysis of covariance), but P-BBMV-GlcNAc is significantly different to all other conditions (P<.5). The line for P-BBMV-GalNAc was not a good fit (P=.7) but was still used for these comparisons. al., 1994). Using similar blotting systems Cry1Ac has been previously reported to bind to proteins of 12 kda (Knight et al., 1994; Knowles et al., 1991) or of both 12 and 21 kda (de Maagd et al., 1996; Martínez-Ramírez et al., 1994) in total M. sexta midgut BBMV preparations. Cry1Ab also binds a 21 kda M. sexta protein (de Maagd et al., 1996; Martínez-Ramírez et al., 1994; Vadlamudi et al., 1993). This Cry1Ab binding component has been identified as a cadherin-like glycoprotein (Vadlamudi et al., 1995). The Cry1Ac binding component observed by us in A-BBMV after an extended blot development (see Results) may correspond to the previously reported 21 kda binding protein (de Maagd et al., 1996; Martínez- Ramírez et al., 1994). However, in this study it was not observed in P-BBMV and therefore cannot explain the proposed second Cry1Ac permeabilising mechanism common to both A- and P-BBMV. The demonstration of the 12 kda binding protein, together with the P-BBMV localisation of aminopeptidase N and the GalNAc inhibition data, suggest that Cry1Ac action in the M. sexta posterior midgut is at least partially mediated via an interaction with aminopeptidase N. However, in the anterior midgut another mechanism must be operating and by inference this mechanism may also be operating in the posterior region in addition to an aminopeptidase N-mediated mechanism. The nature of this second binding component has not been resolved by this study. In addition, it is not known whether both of the Cry1Ac permeabilising activities or only the faster P-BBMV permeability changes are significant in vivo. Cry1Ac binding to anterior regions of the M. sexta midgut has been reported (Bravo et al., 1992a) but that study used isolated midguts to

5 Toxin-induced membrane permeabilisation 313 activation protocol reported to favour channel formation in lipid bilayers was not used (Smedley et al., 1997) and the toxin concentration used was between.5-39 nm, with significant activities being observed at the lower end of this range. Therefore neither of these factors should account for the Cry1Ac interaction with A-BBMV. B. thuringiensis Cry toxins have been shown to bind to glycolipid fractions from a non-target insect (Dennis et al., 1986), and a recent preliminary report has extended this to show Cry toxin interactions with midgut BBMV glycolipids from a target insect M. sexta (Garczynski and Adang, 1996). Conceivably, therefore, membrane lipids may mediate the second Cry1Ac activity that we observed in the M. sexta midgut. If this is the case our data suggest that GalNAc would not be an important component of such a Cry1Ac glycolipid receptor. Fig. 5. Detection of Cry1Ac toxin binding proteins in Manduca sexta midgut A- and P-BBMV. (A) Coomassie blue-stained SDS-13%- polyacrylamide gel. (B) Nitrocellulose paper after transfer of gel proteins, followed by Cry1Ac toxin incubation and subsequent antibody detection of Cry1Ac binding proteins as described in Materials and methods. Lane 1, molecular mass markers; lane 2, A- BBMV, 25 µg; lane 3, P-BBMV, 25 µg. which Cry1Ac was added. Examining the distribution of toxin in the midgut after feeding larvae with Cry1Ac, as was carried out with Cry1Ab (Bravo et al., 1992b), might reveal whether Cry1Ac binds to the anterior midgut brush border membrane in vivo. While quantitative binding studies have demonstrated that in some instances both high and low affinity binding sites are observed for a single Cry toxin-bbmv interaction (Van Rie et al., 199), in the case of Cry1Ac and M. sexta BBMV only a single high affinity binding site has been reported (Garczynski et al., 1991; Van Rie et al., 1989). This suggests that a nonsaturable binding event may mediate one of the two Cry1Ac interactions reported here, which results in an increased M. sexta BBMV permeability. English et al. (1994) reported that the Cry2A δ-endotoxin exhibited only non-saturable binding to BBMV prepared from the midgut of the susceptible insect Helicoverpa zea, whereas the active Cry1Ac toxin demonstrated saturable binding to the same BBMV (English et al., 1994). Interestingly, Cry1Ac was able to compete with the nonsaturable Cry2A binding. Cry toxin channel formation in lipid bilayers in the absence of receptors has been reported (Grochulski et al., 1995; Schwartz et al., 1993; Slatin et al., 199). This activity in lipid bilayers may be the result of a particular toxin activation regime (Smedley et al., 1997) or because the relatively high toxin concentration used favoured channel formation. Direct comparisons between investigations are difficult because of different experimental conditions. However, single studies have reported that the toxin concentration needed to induce pores was markedly reduced when BBMV material was fused with lipid bilayers; for example nm compared with >5 µm needed for pure lipid bilayers (Lorence et al., 1995; Martin and Wolfersberger, 1995). In the present study the toxin We thank Trevor Sawyer for technical assistance and Mr R. Summers and Ms K. Rowsell for photographic work. This work was initiated with support from NIH Grant GM (M.G.W.) and was further supported by a grant from the Biotechnology and Biological Sciences Research Council, UK (J.C. and D.J.E.). REFERENCES Belfiore, C. J., Vadlamudi, R. K., Osman, Y. A. and Bulla, L. A., Jr. (1994). A specific binding protein from Tenebrio molitor for the insecticidal toxin of Bacillus thuringiensis subsp. tenebrionis. Biochem. Biophys. Res. Comm. 2, Bell, R. A. and Joachim, F. G. (1976). Techniques for rearing laboratory colonies of tobacco hornworms and pink bollworms. Ann. Entomol. Soc. Am. 69, Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, Bravo, A., Hendrickx, K., Jansens, S. and Peferoen, M. (1992a). Immunocytochemical analysis of specific binding of Bacillus thuringiensis insecticidal crystal proteins to lepidopteran and coleopteran midgut membranes. J. Invertebr. Pathol. 6, Bravo, A., Jansens, S. and Peferoen, M. (1992b). Immunocytochemical localization of Bacillus thuringiensis insecticidal crystal proteins in intoxicated insects. J. Invertebr. Pathol. 6, Carroll, J. and Ellar, D. J. (1993). An analysis of Bacillus thuringiensis δ- endotoxin action on insect-midgut-membrane permeability using a lightscattering assay. Eur. J. Biochem. 214, Carroll, J. and Ellar, D. J. (1997). Analysis of the large aqueous pores produced by a Bacillus thuringiensis protein insecticide in Manduca sexta midgut-brush-border-membrane vesicles. Eur. J. Biochem. 245, Dennis, R. D., Wiegandt, H., Haustein, D., Knowles, B. H. and Ellar, D. J. (1986). Thin layer chromatography overlay technique in the analysis of the binding of the solubilized protoxin of Bacillus thuringiensis var. kurstaki to an insect glycosphingolipid of known structure. Biomed. Chromatography 1, English, L., Robbins, H. L., Von Tersch, M. A., Kulesza, C. A., Ave, D., Coyle, D., Jany, C. S. and Slatin S. L. (1994). Mode of action of CryIIA: a Bacillus thuringiensis delta-endotoxin. Insect Biochem. Molec. Biol. 24, Feitelson, J. S., Payne J. and Kim L. (1992). Bacillus thuringiensis: Insects and beyond. Bio/Technology 1, Garczynski, S. F. and Adang M. J. (1996). Interactions of Bacillus thuringiensis Cry1 toxins with lipids isolated from the midgut of Manduca sexta larvae. In Abstracts of the 29th Society for Invertebrate Pathology Annual Meeting and IIIrd International Colloquium on Bacillus thuringiensis, p. 28. Society for Invertebrate Pathology, Cordoba, Spain. Garczynski, S. F., Crim J. W. and Adang M. J. (1991). Identification of putative insect brush border membrane-binding molecules specific to Bacillus thuringiensis δ-endotoxin by protein blot analysis. Appl. Environ. Microbiol. 57, Gill, S. S., Cowles, E. A. and Francis V. (1995). Identification, isolation, and cloning of a Bacillus thuringiensis CryIAc toxin-binding protein from the

6 314 J. Carroll, M. G. Wolfersberger and D. J. Ellar midgut of the lepidopteran insect Heliothis virescens. J. Biol. Chem. 27, Grochulski, P., Masson, L., Borisova, S., Pusztai-Carey, M., Schwartz, J. L., Brousseau, R. and Cygler M. (1995). Bacillus thuringiensis CryIA(a) insecticidal toxin: crystal structure and channel formation. J. Mol. Biol. 254, Hafkensheid, J. C. M. (1984). Aminopeptidases and amino acid arylamidases. In Methods of Enzyme Analysis, Vol. V, Enzymes 3; Peptidases, Proteinases and their Inhibitors (ed. H. U. Bergmeyer), pp Verlag Chemie, Weinheim. Hawkes, R., Niday, E. and Gordon, J. (1982). A dot-immunoblotting assay for monoclonal and other antibodies. Anal. Biochem. 119, Hofmann, C., Vanderbruggen, H., Höfte, H., Van Rie, J., Jansens, S. and Van Mellaert, H. (1988). Specificity of Bacillus thuringiensis δ-endotoxins is correlated with the presence of high-affinity binding sites in the brush border membrane of target insect midguts. Proc. Nat. Acad. Sci. USA 85, Höfte, H. and Whiteley, H. R. (1989). Insecticidal crystal proteins of Bacillus thuringiensis. Microbiol. Rev. 53, Knight, P. J. K., Crickmore, N. and Ellar, D. J. (1994). The receptor for Bacillus thuringiensis CryIA(c) delta-endotoxin in the brush border membrane of the lepidopteran Manduca sexta is aminopeptidase N. Mol. Microbiol. 11, Knowles, B. H. (1994). Mechanism of action of Bacillus thuringiensis δ- endotoxins. Adv. Insect Physiol. 24, Knowles, B. H. and Dow, J. A. T. (1993). The crystal δ-endotoxins of Bacillus thuringiensis: models for their mechanism of action on the insect gut. BioEssays 15, Knowles, B. H. and Ellar, D. J. (1987). Colloid-osmotic lysis is a general feature of the mechanism of action of Bacillus thuringiensis δ-endotoxins with different insect specificity. Biochim. Biophys. Acta 924, Knowles, B. H., Knight, P. J. K. and Ellar, D. J. (1991). N-acetyl galactosamine is part of the receptor in insect gut epithelia that recognizes an insecticidal protein from Bacillus thuringiensis. Proc. R. Soc. Lond. B 245, Laemmli, U. K. and Favre, M. (1973). Maturation of the head of bacteriophage T4. I. DNA packaging events. J. Mol. Biol. 8, Lorence, A., Darszon, A., Díaz, C., Liévano, A., Quintero, R. and Bravo, A. (1995). δ-endotoxins induce cation channels in Spodoptera frugiperda brush border membranes in suspension and in planar bilayers. FEBS Lett. 36, Lowry, O. H., Rosebrough, N. J., Farr, A. L. and Randall, R. J. (1951). Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193, Luo, K., Tabashnik, B. E. and Adang, M. J. (1997). Binding of Bacillus thuringiensis Cry1Ac toxin to aminopeptidase in susceptible and resistant diamondback moths. Appl. Environ. Microbiol. 63, de Maagd, R., van der Klei, H., Bakker, P. L., Stiekema, W. J. and Bosch, D. (1996). Different domains of Bacillus thuringiensis δ-endotoxins can bind to insect midgut membrane proteins on ligand blots. Appl. Environ. Microbiol. 62, Martin, F. G. and Wolfersberger, M. G. (1995). Bacillus thuringiensis δ- endotoxin and larval Manduca sexta midgut brush-border membrane vesicles act synergistically to cause very large increases in the conductance of planar lipid bilayers. J. Exp. Biol. 198, Martínez-Ramírez, A. C., González-Nebauer, S., Escriche, B. and Real, M. D. (1994). Ligand blot identification of a Manduca sexta midgut binding protein specific to three Bacillus thuringiensis CryIA-type ICPs. Biochem. Biophys. Res. Comm. 21, Masson, L., Lu, Y-J., Mazza, A., Brousseau, R. and Adang, M. J. (1995). The CryIA(c) receptor purified from Manduca sexta displays multiple specificities. J. Biol. Chem. 27, Meissner, G. (1988). Ionic permeability of isolated muscle sarcoplasmic reticulum and liver endoplasmic reticulum vesicles. Meth. Enzymol. 157, Oddou, P., Hartmann, H. and Geiser, M. (1991). Identification and characterization of Heliothis virescens midgut membrane proteins binding Bacillus thuringiensis δ-endotoxins. Eur. J. Biochem. 22, Sanchis, V. and Ellar, D. J. (1993). Identification and partial purification of a Bacillus thuringiensis CryIC δ-endotoxin binding protein from Spodoptera littoralis gut membranes. FEBS Lett. 316, Sangadala, S., Walters, F. S., English, L. H. and Adang, M. J. (1994). A mixture of Manduca sexta aminopeptidase and phosphatase enhances Bacillus thuringiensis insecticidal CryIA(c) toxin binding and 86 Rb + -K + efflux in vitro. J. Biol. Chem. 269, Schwartz, J. L., Garneau, L., Savaria, D., Masson, L., Brousseau, R. and Rousseau, E. (1993). Lepidopteran-specific crystal toxins from Bacillus thuringiensis form cation- and anion-selective channels in planar lipid bilayers. J. Membr. Biol. 132, Slatin, S. L., Abrams, C. K. and English, L. (199). Delta-endotoxins form cation-selective channels in planar lipid bilayers. Biochem. Biophys. Res. Comm. 169, Smedley, D. P., Armstrong, G. and Ellar, D. J. (1997). Channel activity caused by a Bacillus thuringiensis δ-endotoxin depends on the method of activation. Mol. Membr. Biol. 14, Stewart, G. S. A. B., Johnstone, K., Hagelberg, E. and Ellar, D. J. (1981). Commitment of bacterial spores to germinate. Biochem. J. 198, Thomas, W. E. and Ellar D. J. (1983). Bacillus thuringiensis var. israelensis crystal δ-endotoxin: effects on insect and mammalian cells in vitro and in vivo. J. Cell Sci. 6, Vadlamudi, R. K., Ji, T. H. and Bulla, L. A., Jr. (1993). A specific binding protein from Manduca sexta for the insecticidal toxin of Bacillus thuringiensis subsp. berliner. J. Biol. Chem. 268, Vadlamudi, R. K., Weber, E., Ji, I., Ji, T. H. and Bulla, L. A., Jr. (1995). Cloning and expression of a receptor for an insecticidal toxin of Bacillus thuringiensis. J. Biol. Chem. 27, Valaitis, A. P., Lee, M. K., Rajamohan, F. and Dean, D. H. (1995). Brush border membrane aminopeptidase-n in the midgut of the gypsy moth serves as the receptor for the CryIA(c) δ-endotoxin of Bacillus thuringiensis. Insect Biochem. Mol. Biol. 25, Van Rie, J., Jansens, S., Höfte, H., Degheele, D. and Van Mellaert, H. (1989). Specificity of Bacillus thuringiensis δ-endotoxins. Eur. J. Biochem. 186, Van Rie, J., Jansens, S., Höfte, H., Degheele, D. and Van Mellaert, H. (199). Receptors on the brush border membrane of the insect midgut as determinants of the specificity of Bacillus thuringiensis delta-endotoxins. Appl. Environ. Microbiol. 56, Wiggans, G. N., Andrews, J. S. and Sahgal, A. (1983). COMPREG: a BASIC program to test for differences between several linear regression lines. Behav. Res. Methods Instrum. 15, Wolfersberger, M. G. (1996). Localization of amino acid absorption systems in larval midgut of the tobacco hornworm Manduca sexta. J. Insect Physiol. 42, Wolfersberger, M. G., Lüthy, P., Maurer, A., Parenti, P., Sacchi, V. F., Giordana, B. and Hanozet, G. M. (1987). Preparation and partial characterization of amino acid transporting brush border membrane vesicles from the larval midgut of the cabbage butterfly (Pieris brassicae). Comp. Biochem. Physiol. 86A, Accepted 1 October 1997