Binding analyses with denatured epithelial membrane proteins from Bt (Bacillus thuringiensis) demonstrated at least two kinds of proteins, APNs (aminopeptidases N) and cadherin-like proteins, as possible receptors for the Cry1A class of Bt toxins. Two alternative models have been proposed, both based on initial toxin binding to a cadherin-like protein, but one involving APN and the other not. We have used two Bombyx mori strains (J65 and Kin), which are highly susceptible to Cry1Ab, to study the role of these two types of receptors on Cry1Ab toxin binding and cytotoxicity by means of the inhibitory effect of antibodies. BBMVs (brush-border membrane vesicles) of strain J65 incubated with labelled 125I-Cry1Ab revealed a marked reduction in reversible and irreversible binding when anti-BtR175 (a cadherin-like protein) was used for BBMV pre-treatment. By contrast, the anti-APN1 antibody specifically affected the irreversible binding, while the reversible binding component was not affected. This is the first time that binding of Cry1Ab to APN1 and to a cadherin-like protein from BBMVs in solution has been shown. Dissociated epithelial cells from the Kin strain were used to test the inhibitory effect of the antibodies on the cytotoxicity of Cry1Ab. Pre-incubation of the cells with the anti-BtR175 antibody conferred protection against Cry1Ab, but not the anti-APN1 antibody. Therefore our results seem to support the two models of the mode of action of Cry1Ab in Lepidoptera, depending on whether BBMVs or intact dissociated cells are used, suggesting that both pathways may co-operate for the toxicity of Cry1A toxins in vivo.
Bt (Bacillus thuringiensis) strains produce specific, highly toxic proteins at the moment of sporulation . These proteins accumulate in parasporal aggregations called crystals. The component proteins are called Bt toxins, crystal proteins, Cry proteins or Cry toxins. Hundreds of such proteins  have been identified with activities towards different insects. In spite of the many studies with crystal proteins from the Cry1A subclass (Cry1Aa, Cry1Ab and Cry1Ac), the detailed mode of action of these insecticidal toxins remains unclear . Basically, the larva ingests the toxin in a compact crystalline form (protoxin). The toxin is solubilized in the gut and activated by midgut proteases. The active fragment specifically binds to membrane proteins of the brush border of midgut cells eventually causing cell lysis and the death of the insect. There is evidence that toxin binding involves a two-step interaction with the epithelial membrane, with a reversible component and an irreversible one [4–6].
Binding of the toxin to specific midgut receptors is considered a key step in the toxicity of Cry1 type toxins . In fact, a marked reduction in toxin binding to BBMVs (brush-border membrane vesicles) in solution is considered the most common basis of resistance to these toxins. Binding techniques using denatured proteins, such as ligand blots, have revealed at least two groups of proteins as possible functional Cry1A toxin receptors in different lepidopterans: APNs (aminopeptidases N; EC 126.96.36.199) and a cadherin-like protein. Depending on the insect species, a different set of APN proteins has been proposed to function as Cry1A receptors, such as in the case of Manduca sexta [10–12] or of Bombyx mori [13–15]. However, only one kind of cadherin-like protein has been described as a putative Cry1A toxin receptor in all insect species [16,17]. Numerous experiments have shown the possible involvement of both kinds of proteins in the toxicity of Cry1A toxins, such as the work of Gill and Ellar , in which the 120 kDa APN of M. sexta was expressed transgenically in Drosophila melanogaster, or that of Tsuda et al.  who transfected mammalian cells with the BtR175 gene of B. mori. These studies showed that Cry1A-tolerant insects and mammalian cells respectively became susceptible to Cry1A toxins after transgenesis.
Despite the many binding studies using denatured APNs and cadherin-like proteins, very few have been done on the involvement of these proteins in Cry1A binding while still associated with their native membranes. It is worth noting the inconsistencies found when comparing binding assays using native membranes and ligand blots [9,20]. Indeed, the role the above proteins play in toxin binding and especially in the events that occur following binding is still unclear. One model proposes that the Cry1Ab toxin first binds to the cadherin-like protein and, after further proteolytic activation and oligomerization, it is transferred to an APN receptor, which facilitates the insertion into the membrane with concomitant formation of pores, finally leading to cell death . An alternative model has recently been proposed in which, after Cry1Ab binding to the cadherin-like receptor, a protein kinase A-dependent pathway is activated leading to cell death . In this model, binding to APN and pore formation is irrelevant or of marginal relevance.
To shed light on the role of lepidopteran APNs and the cadherin-like protein in Cry1Ab binding, we used a model lepidopteran species, B. mori. Because of the low toxicity of this toxin to most strains of this insect species (http://www.glfc.scf.rncan.gc.ca/bacillus), we chose two strains (J65 and Kin) that are highly susceptible to Cry1Ab (M.S. Ibiza-Palacios, J. Ferré, S. Higurashi, R. Sato and B. Escriche, unpublished work). Highly specific antibodies against APN1, APN3 and the cadherin-like protein (BtR175) were used to try to inhibit binding of Cry1Ab to the above candidate receptor proteins and thus determine their role in the whole binding process.
Preparation of Bt toxins
Cry1Aa and Cry1Ab protoxins used for larval toxicity bioassays were prepared as follows. Cry1Aa protoxin was prepared from a culture of B. thuringiensis serovar sotto PAS strain. Bacteria were cultured on nutrient agar at 25 °C for 2 days. The crystalline inclusions consisted of Cry1Aa2 protoxin . Cry1Ab protoxin was prepared from a culture of recombinant Escherichia coli expressing the cry1Ab gene . The inclusions that contained Cry1Ab protoxin were partially purified using the method described by Lee et al. . Inclusions of Cry1Aa and Cry1Ab were purified using NaBr-density-gradient centrifugation . Purified inclusions were suspended in sterile distilled water and stored at 4 °C until use. The concentration was estimated using the Lowry assay .
Cry1Ab toxin for the ex vivo midgut toxicity assay was expressed from recombinant E. coli expressing the cry1Ab8 gene from B. thuringiensis serovar aizawai IPL7 strain, which was provided by Sumitomo Chemical. E. coli was grown at 37 °C in 250 ml of LB (Luria–Bertani) medium [1% bacto-tryptone, 0.5% yeast extract and 1% NaCl, supplemented with 50 μg/ml ampicillin and 25 mM IPTG (isopropyl β-D-thiogalactoside)] for 16 h and centrifuged at 7000 g for 35 min. The Cry1Ab protoxin was expressed as an insoluble protein. After incubation in 10 ml of lysis buffer (15 mM Tris/HCl, pH 8.0, 8 mM EDTA and 0.4 mg/ml lysozyme) overnight at 4 °C, bacteria were disrupted by sonication, and inclusion bodies were separated from cell debris by repeating the centrifugation and washing with 5% (w/w) Triton X-100 in PBS (10 mM sodium phosphate buffer, pH 7.4, containing 145 mM NaCl). The inclusion bodies were solubilized and trypsinized, and then the activated toxins were purified using DEAE-cellulose, as described elsewhere . The Bradford method  was used to determine Cry1Ab toxin concentration.
Cry1Ab for the binding analysis was prepared as described by Sayyed et al. . Cry1Ab protoxin was expressed as inclusion bodies in E. coli (kindly supplied by Dr Ruud A. de Maagd, Plant Research International, Wageningen, The Netherlands). Protoxin was solubilized and trypsin-activated. It was purified with a MonoQ HR 5/5 anion-exchange column (Fast Protein Liquid Chromatography system; Amersham, Uppsala, Sweden). The protein concentration was determined by densitometry after SDS/PAGE.
BSA was used as the standard in all protein quantification methods.
Preparation of BBMVs
The Japanese strain No. 65 (J65) of B. mori was obtained from the National Institute of Agrobiological Sciences Gene Bank (Tsukuba, Japan). Insects were reared on artificial diet (Silkmate, Nihon-Nosanko, Japan) at 25 °C. The midguts of fifth instar larvae were freeze-dried as described by Hernandez et al.  and sent to Valencia (Spain), where BBMVs were prepared using the method of Mg2+ precipitation described by Sayyed et al. . The protein concentration was determined using the Bradford method  with BSA as the standard. Two batches of BBMVs were used for all the binding experiments, one for each replicate.
Preparation of antibodies and antiserum
Purified antibodies against GST (glutathione transferase) or fusion proteins of GST with BtR175 and APN1, and antiserum specific to APN3, were prepared by immunizing mice as described in previous reports [15,30]. For immunoblotting, goat anti-mouse IgG–HRP (horseradish peroxidase) conjugate (Bio-Rad) was used as the secondary antibody.
For the larval toxicity assay, suspensions of protoxin inclusions were diluted with sterile distilled water. A four-step serial dilution was made for each protoxin preparation. Mulberry leaf (4 cm2) was smeared with 160 μl of diluted protoxin solution and air-dried before use. The air-dried leaf was placed in a plastic Petri dish (90 mm in diameter), which was covered with filter paper (Advantest No. 2 filter paper), 70 mm in diameter, dampened with a 0.5 ml drop of distilled water to keep the moisture. We used 21–30 newly hatched larvae of strain J65 per piece of leaf. At the end of the second day, the larvae were fed non-treated fresh leaves. Mortality was scored at approx. 48 h intervals for 4 days. The bioassay was performed at 25 °C. Mortality data were analysed using SPSS v. 7.5 software, and the median lethal concentrations (LC50) of the two protoxins were estimated.
For the ex vivo midgut toxicity assay, midguts were dissected from second instar larvae of B. mori strain J65, inverted inside out, cut into 2-mm blocks and incubated at 25 °C in physiological saline (0.9% NaCl) in a 96-well plate. Cry1Ab activated toxin was added, and cell swelling was observed under phase-contrast microscopy.
For the cytotoxicity experiments, the Japanese B. mori strain Kinshu X Showa (Kin) was used. This is a hybrid strain and was purchased from Ueda-Sanshu Co. (Ueda, Japan). Cells were dissociated from dissected midguts as described previously . Cell lysis was assessed by measuring the release of cytosolic LDH (lactate dehydrogenase) . Cells from 4th instar larvae of the Kin strain were pre-incubated for 1 h with 1 μM of either anti-BtR175, anti-APN1, anti-GTS or a mixture of anti-BtR175 and anti-APN1 antibodies. Then, 10 nM Cry1Ab was added and incubated for 1.5 h before measuring the release of LDH activity. The experiment was repeated four times. Control experiments of complete cell lysis were performed using 1% Triton X-100.
All binding experiments were performed within 1 month. Association binding experiments were performed as described by Van Rie et al. . Association and dissociation binding experiments were carried out with 125I-labelled Cry1Ab toxin using the chloramine-T (sodium N-chloro-p-toluenesulfonamide) method . The specific radioactivity of 125I-Cry1Ab was 5.1 mCi/mg, as determined using an ELISA . Association and dissociation binding experiments with antibody pre-incubation were carried out with another batch of Cry1Ab labelled with a specific radio-activity of 3.7 mCi/mg. All binding experiments were performed at least in duplicate.
The association binding assay was performed with different incubation times (0–100 min) and 0.2 mg/ml of BBMVs with 0.17 nM 125I-labelled Cry1Ab for total binding, and with 3 μM (1000-fold) unlabelled Cry1Ab for non-specific binding, in 0.1 ml of binding buffer (PBS buffer: 1 mM KH2PO4, 10 mM Na2HPO4, 137 mM NaCl and 2.7 mM KCl, pH 7.4, with 0.1% BSA) at room temperature (25 °C). Bound toxin was separated from free toxin by centrifugation at 16100 g for 10 min at 4 °C. The pellet was washed twice with 0.5 ml of ice-cold binding buffer. The radio-activity in the pellet was measured in a model 1282 Compugamma CS gamma counter (LKB Pharmacia). Specific binding was calculated by subtracting non-specific binding from total binding.
For the dissociation experiments, an excess (1000-fold) of unlabelled toxin was added after 100 min of association binding. After different incubation times (0–125 min), the reaction mixture was centrifuged at 16100 g for 10 min at 4 °C. Control experiments followed the same protocol, but PBS buffer was added instead of unlabelled toxin. The non-specific binding was determined by incubating BBMVs with both labelled and excess (1000-fold) unlabelled Cry1Ab for 100 min. The value of the specific binding was obtained from the association binding assay.
Antibody inhibition of each protein was assessed using a 1 h pre-incubation of the BBMVs used in the binding assays with 1 μM of specific antibody or a 1:20000 dilution of antiserum in binding buffer at room temperature. Then, the dissociation experiment protocol was followed, as described above, but using 0.34 nM 125I-Cry1Ab. The total specific binding was obtained at 100 min from the association binding experiment with the antibody-pre-incubated BBMV. The integrity of the vesicles was determined by comparing the values with a control experiment using PBS buffer instead of the specific antibody when pre-incubated with the BBMV. The stability of the binding association was checked by adding PBS buffer instead of unlabelled toxin. The irreversible binding was measured as the specific binding remaining in a dissociation experiment after 100 min. Reversible binding was calculated as the difference between total specific binding and irreversible binding.
Susceptibility of the strain J65 to Cry1A toxins
The susceptibility of B. mori strain J65 to Cry1Aa and Cry1Ab toxins was evaluated using two methods, a larval mortality assay and a cytological assay. The two toxins showed similar toxicities to neonatal larvae based on the mortality after 4 days (Table 1). However, although LC50 values were similar, the larval mortality responses cannot be considered equal as the slopes were significantly different (P=0.04, t test), indicating that the response of strain J65 to increasing concentrations of Cry1Aa was faster than the response to Cry1Ab.
Because many strains of B. mori are very susceptible to Cry1Aa but not to Cry1Ab, the unusually high susceptibility of B. mori strain J65 to the Cry1Ab toxin was confirmed by an ex vivo bioassay. Dissected midguts were incubated with two concentrations of toxin (1 and 10 nM) for different periods (1, 2 and 3 h). Microscopic observation of the toxin-exposed midguts revealed the typical effects of Cry proteins , including the swelling of midgut cells (Figure 1). This effect was dependent on the dose of toxin and incubation time, demonstrating that Cry1Ab is toxic to the midgut cells. Control midguts were either not treated with toxin or were exposed for only 1 min; in both cases, no effect was observed (Figure 1).
Ex vivo susceptibility of midgut epithelial cells from the B. mori strain J65 to Cry1Ab
Kinetics of Cry1Ab binding
Representative kinetics of the binding of Cry1Ab in solution to BBMVs from larval midguts of strain J65 insects is shown in Figure 2. The reversible and irreversible components of the total binding were estimated from an association experiment to measure the total specific binding followed by a dissociation experiment to determine the amount of irreversible binding. BBMVs incubated with 0.17 nM 125I-Cry1Ab showed stable specific binding after 100 min (at approx. 6% of added ligand). The subsequent addition of an excess of unlabelled toxin decreased the amount of specific binding, which stabilized after 125 min at approx. 1.5% of added ligand. This 1.5% represents the irreversibly bound fraction, which was approx. 25% of the total specific binding that occurred in the association experiment. The difference between the total binding and the irreversible binding (i.e. 4.5% of added ligand) represents the reversible component of the binding and accounted for approx. 75% of the total specific binding. The non-specific binding, determined by adding an excess of unlabelled toxin to the initial mixture, was approx. 2.5±0.1% (means±S.D.) of the added ligand. This value remained stable regardless of the length of the experiment. Similarly, control experiments with no addition of unlabelled toxin after the association kinetics experiment showed that the amount of total specific binding was maintained constant for at least 125 min.
Association and dissociation kinetics of the binding of 125I-Cry1Ab to larval BBMVs from B. mori strain J65
Antiserum and antibodies specificity
The specificity of the antiserum raised against APN3 and the antibodies raised against the APN1 and BtR175 proteins of B. mori was tested by Western-blot analysis with BBMV proteins from the J65 strain (Figure 3). The observed 115 kDa APN1 band and 110 kDa APN3 band agreed with a previous report . In our hands, the anti-APN1 antibody also recognized a 110 kDa band which could represent some cross-reactivity of APN3. However, we can exclude this possibility because, under the same experimental conditions, the anti-APN1 antibody recognized only a 115 kDa band in another strain of B. mori that possesses APN3 (M.S. Ibiza-Palacios, J. Ferré, S. Higurashi, R. Sato and B. Escriche, unpublished work). It is possible that the 110 kDa protein was a breakdown product of APN1. The 195 kDa band is presumably BtR175, as it was recognized by the anti-BtR175 antibody and its size was similar to the 195 kDa band seen in the middle and posterior midgut regions of B. mori on immunoblotting . Nagamatsu et al.  also purified this 195 kDa band from the BtR175 fraction and visualized it by immunoblotting with anti-BtR175 antibodies. Control experiments performed with antibodies raised against the GST protein did not recognize any band. The results suggest that each antibody specifically interacted with the single class of protein against which it was raised.
SDS/PAGE and Western-blot analysis of BBMV proteins from B. mori strain J65
Inhibition of Cry1Ab binding using specific antibodies
In order to determine which membrane proteins were involved in the binding of Cry1Ab, BBMVs from the J65 strain were pre-incubated for 1 h with antibodies against candidate receptors before starting the binding assays. Significant changes in the binding kinetics after incubation with the antibodies were interpreted as the participation of the specific protein recognized by the antibody in the binding process.
In relation to the control with just buffer, no significant effect (one-way ANOVA with the Bonferroni post-test, P>0.05) on total specific binding was observed in samples pre-incubated with either anti-APN3 serum or anti-GST antibody (Figure 4). However, significant reduction in total specific binding (one-way ANOVA with the Bonferroni post-test, P<0.05) was observed when anti-BtR175 and anti-APN1 antibodies were used. Anti-BtR175 produced a greater reduction (84%) than anti-APN1 (55%). Control experiments were performed to determine the possible effect of the antibody incubation on the non-specific binding; no variation in this component of binding was observed.
Association binding experiments of 125I-Cry1Ab and BBMVs from B. mori J65
Analysis of the effect of antibody pre-incubation on the reversible and irreversible components of binding revealed a marked reduction of both components when using anti-BtR175 (Figure 5), with values not significantly different (t test, P>0.05) from the values obtained for non-specific binding. By contrast, anti-APN1 specifically affected the amount of irreversible binding (to the same level as the non-specific binding) (Figure 5A), while the reversible binding component was not affected (t test, P>0.05) (Figure 5B).
Dissociation binding experiments of 125I-Cry1Ab and BBMVs from B. mori J65 strain
Inhibition of Cry1Ab cytotoxicity to dissociated midgut cells
The effect of anti-APN1 and anti-BtR175 antibodies on the action of Cry1Ab on intact cells (as opposed to BBMV) was measured using the Kin strain, another B. mori Cry1Ab-susceptible strain (M.S. Ibiza-Palacios, J. Ferré, S. Higurashi, R. Sato and B. Escriche, unpublished work). The release of cytosolic LDH (as an indicator of cell lysis) upon incubation of dissociated cells with Cry1Ab demonstrated the cytotoxicity of this protein ex vivo (Figure 6). Pre-incubation of the cells with anti-BtR175 antibody significantly reduced the cytotoxic effect. However, neither anti-APN1 nor anti-GTS (control) antibodies significantly decreased the cytotoxicity of Cry1Ab under the assay conditions. In addition, anti-APN1 did not show any additive effect when combined with the anti-BtR175 antibody.
Effect of anti-BtR175 and anti-APN1 antibodies on the cytotoxicity of Cry1Ab on B. mori dissociated midgut epithelial cells as measured by the release of cytosolic LDH
B. mori has been shown to be tolerant to Cry1Ab and Cry1Ac proteins and susceptible to Cry1Aa. However, it is not surprising that different strains can show variation in susceptibility, as was found in other lepidopteran species . In fact, the relative toxicities of Cry1Aa, Cry1Ab or Cry1Ac vary markedly in different reports [4,23,34–38]. Strains J65 and Kin are atypical strains in that they are highly susceptible to Cry1Ab (the present study and M.S. Ibiza-Palacios, J. Ferré, S. Higurashi, R. Sato and B. Escriche, unpublished work), to a similar extent as to Cry1Aa. Our experiments analysing several susceptible and resistant B. mori strains indicated that the resistance of several strains does not depend on a loss of toxin-activating activity or an excess of toxin-degrading activity of the midgut fluid (results not shown), but rather on the direct susceptibility of midgut epithelial cells to the toxin (Figure 2).
The association binding kinetics of toxins with BBMVs in solution provide information on the full binding process, but dissociation kinetics experiments are needed to determine the relevance of each binding component, reversible and irreversible. In fact, Liang et al.  showed that the rates of irreversible binding of Cry1A toxins, rather than the binding affinities, were directly related to toxicity in Lymantria dispar BBMV. Other studies have shown that loss of toxicity and decreased irreversible binding without change in reversible binding parameters occur simultaneously by mutating toxins in domain I or II [5,39]. The association of Cry1Ab toxin with membranes from B. mori has been shown previously using BBMVs in solution using a Cry1Ab-tolerant strain . However, with the tolerant strain, the amount of irreversible binding observed was much lower than in our study (8% versus 30%, referred to as the total specific binding). Ihara et al.  proposed a direct relationship between Cry toxin toxicity and irreversible binding in B. mori. Therefore the higher Cry1Ab irreversible binding contribution to the total binding in strain J65 compared with that in the tolerant strain of Ihara et al.  must be responsible for the high Cry1Ab susceptibility of J65.
Our study using specific antibodies to block toxin binding allowed the identification of the molecules involved in the binding process. Denaturing techniques, such as ligand-blot analysis, in several lepidopteran species have determined that Cry1Ab binds to several types of APNs [7,10] and also to cadherin-like proteins, such as Bt-R1 in M. sexta  and OnBt-R1 in Ostrinia nubilalis . In B. mori, Nakanishi et al.  reported binding of Cry1Ab to APN1 but not to BtR175, and no other study has provided data on the binding of Cry1Ab to BtR175 of B. mori. However, our results suggest that Cry1Ab toxin binds to the putative BtR175 receptor in strain J65. This is somewhat expected given that the strain we have used is susceptible to this toxin and that cadherin-like proteins are thought to play an important role in the toxic action of Cry1A toxins in lepidopterans [16,19,30,42,43]. An observation that supports our finding is that Nagamatsu et al.  showed the inhibition of Cry1Aa binding to BBMVs of B. mori after incubating the BBMVs with anti-BtR175 antibody, and Cry1Aa and Cry1Ab toxins have been shown to share binding sites in B. mori .
Our results of binding inhibition using antibodies showed that both APN1 and BtR175 are involved in Cry1Ab binding to B. mori BBMVs in solution. The use of anti-BtR175 antibody practically completely abolished Cry1Ab binding, affecting both its reversible and irreversible components. However, the anti-APN1 antibody only inhibited the irreversible component of Cry1Ab binding (Figures 4 and 5). These results can be explained by a model in which binding is the result of sequential steps involving these two proteins, such as the model proposed by Bravo et al. . In this model, the initial binding of Cry1Ab toxin to the cadherin-like protein would be reversible, in order to drive the toxin to the APN to cause pore formation. This last step would require stable irreversible binding to permit the lytic activity. Our results are in agreement with this model in that blocking the first step (the cadherin-like receptor) inhibited the whole binding process, while blocking the second step (the APN1 receptor) only inhibited the irreversible binding component, allowing the reversible binding to the cadherin-like receptor to take place. In addition, our experiments identified APN1 of B. mori as the specific APN involved in this model, because the treatment with anti-APN1 antibody totally blocked the irreversible binding. This is also in agreement with the reported functional interaction of a Cry1A (Cry1Ac) with APN1 in Helicoverpa armigera  and with Bravo et al.  and Gómez et al.  because the 120 kDa APN of M. sexta in the proposed model is homologous with B. mori APN1 . No other APNs or other proteins seem to be significantly involved in binding in B. mori, given that we have been able to inhibit essentially all binding by using highly specific antibodies.
Our results with the antibody inhibition of the cytotoxic action of Cry1Ab are in complete agreement with previous studies in B. mori in which the involvement of APN1 in the cytotoxicity of Cry1Aa or Cry1Ac seemed not to be relevant  in this type of ex vivo assay and that only binding to the BtR175 receptor seemed relevant. In our hands, either the anti-APN antibody cannot prevent efficiently the binding of the toxin in this type of assay or other mechanisms of toxicity are acting in live cells as opposed to BBMV. The latter possibility would be in agreement with the model proposed by Zhang et al. , in which binding to the cadherin-like protein would be sufficient to provoke the death of detached cells. It is possible that both mechanisms (pore formation and the signalling pathway) can act simultaneously in vivo (in the intact larva) and that, depending on the experimental system used, the results can be the consequence of one or the other. In addition, the relative contribution of both types of toxicity pathway can be different depending on the lepidopteran species considered.
Our work provides the first evidence that Cry1Ab toxin binds to APN1 and to BtR175 cadherin-like protein in BBMVs in solution. All the previous lines of evidence showing the binding of Cry1Ab to these receptors have been obtained either under denatured conditions or with purified proteins out of their membrane environment. Furthermore, our results seem to support the two models of the mode of action of Cry1Ab in Lepidoptera. On the one hand, our results are in agreement with the sequential involvement of BtR175 and APN1 in the binding to BBMVs . On the other hand, the assays with dissociated cells are in agreement with the involvement of the cadherin-like receptor in triggering a signalling pathway . It is very likely that both pathways contribute to the toxicity of Cry1A toxins in vivo. If confirmed, it would be desirable that further studies be undertaken to determine the relative contribution of both mechanisms to the in vivo toxic effect of Cry1A toxins.
We thank Dr Juan L. Jurat-Fuentes for useful comments that helped to improve this paper, Dr Ruud A. de Maagd for supplying the Cry1Ab E. coli strain for binding experiments, and the Biotechnology Laboratory, Takarazuka Research Center Sumitomo Chemical Co. Ltd, for kindly providing an E. coli strain expressing Cry1Ab8 toxin. Work in Spain was supported by Generalitat Valenciana grants (GV04B-165 and Grupos2004-21) and by a research contract for B. E. from the ‘Ramón y Cajal’ programme. M. S. I.-P. was supported by the Ministerio de Educación, Cultura y Deportes, Spain with a fellowship (AP 2001-0972).