INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Use of insecticides derived from the bacterium Bacillus thuringiensis has increased with commercialization of transgenic plants. Transgenic corn and cotton producing B. thuringiensis toxins are already on the market, and several other B. thuringiensis-transgenic crops could be introduced soon (8, 21). Because insects can adapt to B. thuringiensis toxins, the long-term efficacy of B. thuringiensis toxins in transgenic crops or in sprays will depend on understanding and managing pest resistance to B. thuringiensis toxins (15, 16, 38, 44).

The mode of action of B. thuringiensis toxins involves ingestion followed by crystal solubilization and proteolytic activation of protoxin in the insect midgut. Activated toxin binds to receptors in the midgut epithelial membrane and inserts into the membrane, leading to cell lysis and death of the insect (34). Binding of B. thuringiensis toxins to specific sites in the epithelial membrane is a key step in toxin specificity (20, 45). Reduced binding of toxin to midgut membrane target sites is the best-known mechanism of resistance to B. thuringiensis toxins (4, 14, 19, 33, 47). Furthermore, alteration of a common binding site has been found in cases where insects evolved high levels of resistance simultaneously to more than one B. thuringiensis toxin (28, 33, 39).

Some resistance management strategies rely on sequential or simultaneous use of different B. thuringiensis toxins (32). For such strategies to work, cross-resistance must not occur among the different toxins. So far, the best method of predicting cross-resistance among B. thuringiensis toxins is determining which toxins share a common binding site in a given insect. Sequences or combinations of toxins that share a common binding site are not likely to be useful for managing resistance.

Most lepidopteran insects tested for binding of Cry1Aa, Cry1Ab, and Cry1Ac toxins to midgut brush border membrane vesicles (BBMV) share a common binding site for these toxins (4, 13, 23, 24). This is not surprising, because these three Cry1A toxins have 73 to 88% amino acid sequence identity for the activated toxin (as determined by using the BLAST program [1]).

Cry1Ja is a lepidopteran active toxin (10) with low sequence identity with Cry1A toxins (e.g., 47% amino acid sequence identity with Cry1Ac). Nonetheless, two strains of Plutella xylostella selected with B. thuringiensis products containing Cry1A toxins but not Cry1Ja evolved resistance to Cry1A toxins and strong cross-resistance to Cry1Ja (10, 39). In this pest, one gene confers resistance to Cry1Aa, Cry1Ab, Cry1Ac, Cry1Fa, and Cry1Ja toxins (36, 40). It was shown elsewhere that binding of Cry1Aa, Cry1Ab, and Cry1Ac toxins was strongly reduced in some resistant populations (4, 39). Because Cry1Fa competes for Cry1A's binding site (17), it has been proposed that binding of this toxin might also be affected in the resistant populations.

P. xylostella cross-resistance to Cry1Ja suggests that this toxin might also share a binding site with the Cry1A toxins and Cry1Fa. Recently, a common binding site for Cry1Aa, Cry1Ab, Cry1Ac, Cry1Fa, and Cry1Ja was reported for Heliothis virescens (Noctuidae) (22). Here we used competitive binding experiments to test the hypothesis that Cry1Ja and Cry1Ac share binding sites in six moth and butterfly species, each from a different family: Cacyreus marshalli (Lycaenidae), Lobesia botrana (Tortricidae), Manduca sexta (Sphingidae), Pectinophora gossypiella (Gelechiidae), P. xylostella (Plutellidae), and Spodoptera exigua (Noctuidae). We performed binding competition experiments with biotinylated Cry1Ja and unlabeled Cry1Ac as well as radioactively labeled Cry1Ac and unlabeled Cry1Ja.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Toxin preparation and labeling. Cry1Ab, Cry1Ac, and Cry1Ja toxins employed in the study were prepared from recombinant B. thuringiensis strains expressing a single toxin (strains EG7077, EG11070, and EG7279, respectively). Toxins were purified and activated as described elsewhere (33). For bioassays, toxins were used in their activated form (for C. marshalli, P. xylostella, and S. exigua) or as a lyophilized powder of spores and crystals (for M. sexta, L. botrana, and P. gossypiella). Toxins used for labeling and binding experiments were chromatographically purified using a MonoQ HR 5/5 anion-exchange column (fast protein liquid chromatography system from Pharmacia, Uppsala, Sweden) (33). Protein concentration was determined by the method of Bradford (5) for the activated toxins and by densitometric analyses of the toxin band in gel electrophoresis for the lyophilized powder.

Bioassays. Bioassays were performed in a rearing chamber at 25°C with 60% RH and a 16-h-8-h photoperiod (light-dark). Different bioassay methods were employed depending on the insect tested. For C. marshalli, petals of geranium (Pelargonium × hortorum) were dipped in toxin solutions in 50 mM carbonate buffer (pH 10.5)-0.02% Triton AG-98. Petals were air dried and deposited over 2% agar. First-instar larvae were allowed to feed on these petals, and mortality was scored after 3 days. The surface contamination method of artificial diet was used for L. botrana, M. sexta, and S. exigua. Neonates of M. sexta were tested with lyophilized powder in water, and mortality was scored after 3 days. Third-instar larvae of L. botrana were tested with lyophilized powder in water, and mortality was scored after 5 days. Neonates of S. exigua were tested with toxin solutions in 50 mM carbonate buffer, pH 10.5, and mortality was scored after 5 days. Neonates of P. gossypiella fed on wheat germ diet with toxin incorporated (37). For P. xylostella, we used the leaf dip method described by Tabashnik et al. (41) with third-instar larvae of the LAB-V strain and toxin solutions in 50 mM carbonate buffer, pH 10.5, and mortality was scored after 2 days. Controls with buffer or water without toxin were used in all cases to estimate natural mortality.

Preparation of BBMV. Midguts from last-instar larvae were used to prepare BBMV according to the method described by Wolfersberger et al. (48). In the case of the small insects, such as P. xylostella, L. botrana, and P. gossypiella, whole larvae instead of dissected midguts were used (12). BBMV protein concentration was determined by the method of Bradford (5).

Binding of biotinylated Cry1Ja. Biotinylated Cry1Ja was incubated for 1 h with BBMV in 0.1 ml of binding buffer (phosphate-buffered saline-bovine serum albumin; 8 mM Na₂HPO₄, 2 mM KH₂PO₄, 150 mM NaCl [pH 7.5], and 0.1% bovine serum albumin). BBMV were washed twice with 0.5 ml of binding buffer and resuspended in 10 µl of electrophoresis sample buffer (26). Samples were electrophoresed in a sodium dodecyl sulfate (SDS)-10% polyacrylamide gel and electrotransferred to a nitrocellulose membrane (Hybond-C Super; Amersham). Biotinylated Cry1Ja in the membrane was detected by chemiluminescence according to the manufacturer's instructions in the ECL kit (RP2209; Amersham). Incubations with biotinylated Cry1Ja and at least a 200-fold excess of unlabeled Cry1Ja, Cry1Ac, or Cry1Ab were performed at the same time for each insect. Incubation conditions were adjusted for each insect. For C. marshalli, M. sexta, and P. gossypiella, 20 µg of BBMV was used with 10 ng of biotinylated Cry1Ja. For L. botrana and P. xylostella, incubations were performed with 10 µg of BBMV and 20 ng of biotinylated Cry1Ja. For S. exigua, 30 µg of BBMV was incubated with 10 ng of biotinylated Cry1Ja.

Binding of ¹²⁵I-Cry1Ac and ¹²⁵I-Cry1Ab to BBMV. Binding experiments were performed as described elsewhere (49) using appropriate conditions for each insect regarding incubation time, BBMV concentration, labeled toxin concentration, and dilutions of cold toxin. With M. sexta, L. botrana, and S. exigua, incubation was performed for 1 h using 50 ng of ¹²⁵I-Cry1Ac per ml and 50, 100, and 75 µg of BBMV proteins per ml, respectively. With P. gossypiella, incubation was 50 min with 50 ng of ¹²⁵I-Cry1Ac per ml and 60 µg of BBMV proteins per ml. With C. marshalli, conditions were 45 min and 50 ng of ¹²⁵I-Cry1Ac per ml and 50 µg of BBMV proteins per ml. With P. xylostella, incubations were carried out for 30 min using 50 ng of ¹²⁵I-Cry1Ac per ml and 70 µg of BBMV proteins per ml or 5 ng of ¹²⁵I-Cry1Ab per ml and 100 µg of BBMV proteins per ml. Values for the dissociation constant (K_d) were obtained from competition binding data by the procedure described by Munson and Rodbard (30) with GraphPad software.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Susceptibility of larvae to Cry1Ac and Cry1Ja. The toxicity of Cry1Ac and Cry1Ja varied among the species tested (Table 1). Cry1Ja was less toxic than Cry1Ac for four species (L. botrana, M. sexta, P. gossypiella, and P. xylostella), but the opposite occurred for one species (C. marshalli). Neither Cry1Ac nor Cry1Ja was highly toxic to S. exigua. Maximum mortality of S. exigua was 6.5%, even though the concentration tested was 10 times higher than the concentration that killed 44.5 to 96.0% of L. botrana and M. sexta insects in the same type of bioassay.

Competition between Cry1Ja and Cry1A toxins for BBMV binding sites as determined with biotinylated Cry1Ja. Biotinylated Cry1Ja bound to BBMV from all species tested (Fig. 1). Addition of excess unlabeled Cry1Ja substantially reduced binding of biotinylated Cry1Ja, indicating that most of the Cry1Ja binding was specific. An excess of unlabeled Cry1Ac reduced binding of biotinylated Cry1Ja markedly in five species (L. botrana, M. sexta, P. gossypiella, P. xylostella, and S. exigua) and to a lesser degree in C. marshalli (Fig. 1). These results suggest that Cry1Ja binding sites are shared by Cry1Ac in all six species, with a lower degree of sharing for C. marshalli. For P. xylostella, competition with unlabeled Cry1Ab revealed that the Cry1Ja binding site is also shared with Cry1Ab.

View larger version (51K): [in this window] [in a new window]   FIG. 1.   Binding of biotin-labeled Cry1Ja to BBMV from C. marshalli (a), L. botrana (b), M. sexta (c), P. gossypiella (d), P. xylostella (e), and S. exigua (f), in the absence of competitor (lanes labeled ) or in the presence of a 200-fold excess of competitor (Cry1Ja, Cry1Ac, and Cry1Ab lanes). Biotinylated Cry1Ja was incubated with BBMV and then subjected to SDS-polyacrylamide gel electrophoresis. After transfer to nitrocellulose membranes, biotinylated Cry1Ja was detected by chemiluminescence.

Competition between Cry1Ja and Cry1A toxins for BBMV binding sites as determined with ¹²⁵I-Cry1A toxins. Competition binding experiments with ¹²⁵I-labeled Cry1Ac and unlabeled Cry1Ja indicated that Cry1Ja competed for Cry1Ac binding sites in all of the insects tested (Fig. 2). Cry1Ja competed completely for Cry1Ac-specific binding in C. marshalli and P. gossypiella (Fig. 2A and D) and almost completely in M. sexta and S. exigua (Fig. 2C and F). This indicates that Cry1Ja recognizes most, if not all, binding sites used by Cry1Ac in the aforementioned species. In L. botrana and P. xylostella, only about 50% of the bound ¹²⁵I-Cry1Ac was competed off at the highest concentration of unlabeled Cry1Ja used (Fig. 2B and E). This result indicates that Cry1Ac binds to different types of sites and that Cry1Ja binds to some, but not all, of them.

View larger version (35K): [in this window] [in a new window]   FIG. 2.   Binding of 125I-Cry1Ac to BBMV from C. marshalli (A), L. botrana (B), M. sexta (C), P. gossypiella (D), P. xylostella (E), and S. exigua (F) at different concentrations of nonlabeled Cry1Ac () or Cry1Ja ().

View larger version (18K): [in this window] [in a new window]   FIG. 3.   Binding of 125I-Cry1Ab to BBMV from P. xylostella at different concentrations of nonlabeled Cry1Ab () or Cry1Ja ().

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Despite the large number of B. thuringiensis Cry toxins known, only a handful are currently used to control lepidopteran pests (K. Van Frankenhuyzen and C. Nystrom, Bacillus thuringiensis toxin specificity database [http://www.glfc.forestry.ca/english/res/Bt_HomePage/netintro.htm]). Although Cry1A toxins are the most widely used commercially, Cry1Ja is also effective against a wide variety of lepidopteran pests (10; present study). However, the potential use of Cry1Ja as an alternative to Cry1A toxins is jeopardized because some strains of P. xylostella with resistance to Cry1A toxins are cross-resistant to Cry1Ja (10, 39).

Our results with P. xylostella show that Cry1Ja, Cry1Ab, and Cry1Ac bind to common sites in the brush border membrane of the larval epithelium. In P. xylostella, binding of biotinylated Cry1Ja is almost completely inhibited by an excess of Cry1Ab or Cry1Ac (Fig. 1), but Cry1Ja binds with high affinity only to a small fraction of binding sites used by Cry1Ac and Cry1Ab (Fig. 2E and 3). We propose that an alteration of these high-affinity binding sites in P. xylostella might confer resistance to Cry1Ja as well as to Cry1A toxins.

The bimodal competition curves for Cry1Ja in P. xylostella imply the occurrence of two types of binding sites for Cry1Ac and Cry1Ab. Previous studies had indicated a single binding site for these two toxins in this insect (2, 4, 14). Furthermore, from the homologous competition data from the present work, only a single binding site is evident for Cry1Ac and Cry1Ab. Therefore, Cry1Ac and Cry1Ab must bind with such a similar affinity to different types of binding sites in P. xylostella that analysis of data from homologous competition curves or from heterologous competition curves among Cry1A toxins (2, 4) cannot reveal the occurrence of more than one binding site. However, the use of a competitor (Cry1Ja) with different affinities for two types of binding sites does reveal their existence.

Cry1Ja and Cry1A toxins are potent against species of the family Noctuidae (Van Frankenhuyzen and Nystrom, Bacillus thuringiensis toxin specificity database). However, in S. exigua, both Cry1Ac and Cry1Ja are only marginally toxic, even though they bind with relatively high affinity to the shared sites in BBMV. Low toxicity of Cry1A toxins associated with high-affinity binding has been described previously for S. exigua (13, 29) and Spodoptera frugiperda (29). This observation indicates that mechanisms other than lack of binding, involved in the toxic pathway of Cry toxins, are responsible for the low susceptibility of these species. Luo et al. (29) showed that Cry1Ac did not permeabilize BBMV from the above two Spodoptera species.

For the other four lepidopteran species studied here, Cry1Ja and Cry1Ac also share common binding sites, but the binding patterns vary somewhat among species. In C. marshalli, binding affinity is similar for Cry1Ja and Cry1Ac. However, in L. botrana, M. sexta, and P. gossypiella, Cry1Ja binds with less affinity than does Cry1Ac. Despite the low affinity with which Cry1Ja binds to BBMV from these species, an excess of Cry1Ac completely impedes binding of biotinylated Cry1Ja. Therefore, Cry1Ja binds only to the site detected, and this site is responsible for its toxicity. Cry1Ja did not compete completely with ¹²⁵I-labeled Cry1Ac in L. botrana, which indicates that this insect has more than one binding site for Cry1Ac.

The amino acid sequence identity between Cry1Ac and Cry1Ja is 47% for the activated toxin and 42% for domain II (1). Although domain II and domain III of the Cry toxins may be involved in binding (3, 7, 27), studies with mutant toxins have identified specificity-determinant regions in three loops of domain II (34). Figure 4 shows a comparison of the amino acid sequences of domain II between Cry1A toxins, Cry1Ja, and Cry1Ca. Cry1Ca does not bind to Cry1A binding sites and is included here as a negative control (11, 17, 29, 46). Looking for conserved amino acids within loops in the three Cry1A toxins, we find 3 out of 4 in loop 1, 6 out of 13 in loop 2, and 2 out of 10 in loop 3. However, restricting attention to amino acids conserved in Cry1Ja but not in Cry1Ca leaves just a common Arg (R) in loop 1, a common Gln (Q) or Ser (S) in loop 2, and a common Ser or Gln in loop 3. Moreover, changing the also conserved Ser in loop 3 of Cry1Ac had no effect on toxin binding in M. sexta (35). There are three more amino acids conserved in Cry1Ja and Cry1A toxins, but not in Cry1Ca, just next to loops 2 (Y³⁶⁶) and 3 (R⁴⁴⁷ and A⁴⁴⁹). It is not then evident which amino acids in Cry1Ja are responsible for this toxin competing for binding with Cry1A toxins. It is possible that Cry1A toxins and Cry1Ja utilize, for binding, different overlapping epitopes in the same membrane molecule or even different membrane molecules which are clustered or in close proximity. It is also possible that Cry1A toxins and Cry1Ja interfere with each other's binding through binding by domain III.

View larger version (49K): [in this window] [in a new window]   FIG. 4.   Amino acid sequence alignment for domain II regions of Cry1Aa (M11250), Cry1Ab (M13898), Cry1Ac (M11068), Cry1Ja (L32019), and Cry1Ca (X07518). Alignment was produced with the CLUSTALW program (42). Potential loops were identified with Cry1Aa as a reference point (18). Conserved sequences are boxed. Black boxes indicate positions where amino acid residues are identical. White boxes indicate positions where either at least four amino acid residues are similar or only four residues are identical.

Two kinds of proteins have been characterized as candidate receptors of Cry1A toxins. A cadherin-like protein is recognized by Cry1Ab toxin in M. sexta (43), and aminopeptidase- N (APN) has been described as a Cry1Ac-binding protein in P. xylostella (four different proteins), M. sexta (two different proteins), and other insects (6, 9, 25, 31). Comparison of amino acid sequences from 11 APN putative receptors from different insects has shown a highly conserved region for Cry1Aa binding (31). If APN were one of the membrane proteins used by Cry1A toxins to bind in vivo, it would be possible that this conserved region was the site where Cry1A toxins and Cry1J bind.

In all six species tested here and in H. virescens (22), Cry1Ac and Cry1Ja competed for common binding sites. Because these seven species represent six families, we propose that shared binding sites for Cry1Ac and Cry1Ja are common among Lepidoptera. Evolution of resistance to more than one toxin is associated with the alteration of a common binding site in several insect species (19, 28, 39). Thus, knowledge of which toxins share binding sites can help in choosing appropriate sets of toxins for delaying resistance. Because Cry1Ja and Cry1A toxins share common binding sites in all species of Lepidoptera tested so far, we discourage the combination of Cry1A toxins with Cry1Ja for pest control.