|Home | About | Journals | Submit | Contact Us | Français|
Cry4Aa produced by Bacillus thuringiensis is a dipteran-specific toxin and is of great interest for developing a bioinsecticide to control mosquitoes. Therefore, it is very important to characterize the functional motif of Cry4Aa that is responsible for its mosquitocidal activity. In this study, to characterize a potential receptor binding site, namely, loops 1, 2, and 3 in domain II, we constructed a series of Cry4Aa mutants in which a residue in these three loops was replaced with alanine. A bioassay using Culex pipiens larvae revealed that replacement of some residues affected the mosquitocidal activity of Cry4Aa, but the effect was limited. This finding was partially inconsistent with previous results which suggested that replacement of the Cry4Aa loop 2 results in a significant loss of mosquitocidal activity. Therefore, we constructed additional mutants in which multiple (five or six) residues in loop 2 were replaced with alanine. Although the replacement of multiple residues also resulted in some decrease in mosquitocidal activity, the mutants still showed relatively high activity. Since the insecticidal spectrum of Cry4Aa is specific, Cry4Aa must have a specific receptor on the surface of the target tissue, and loss of binding to the receptor should result in a complete loss of mosquitocidal activity. Our results suggested that, unlike the receptor binding site of the well-characterized molecule Cry1, the receptor binding site of Cry4Aa is different from loops 1, 2, and 3 or that there are multiple binding sites that work cooperatively for receptor binding.
Bacillus thuringiensis subsp. israelensis has received considerable attention for mosquito control because of its specific and potent toxicity (15). B. thuringiensis subsp. israelensis-based microbial insecticides have been widely used as active components for integrated management of mosquitoes (11, 13, 33, 34). B. thuringiensis subsp. israelensis produces at least four major crystal toxins (Cry toxins), namely, Cry4Aa, Cry4Ba, Cry11Aa, and Cyt1Aa (5). Cry4Aa exhibits specific toxicity against Anopheles, Aedes, and Culex mosquito larvae (15, 27). The 130-kDa Cry4Aa protoxin is released from the protein crystal upon ingestion by susceptible mosquito larvae and is activated by gut proteases into two protease-resistant fragments with molecular masses of 20 and 45 kDa through intramolecular cleavage of a 60-kDa intermediate (39). The three-dimensional structure of Cry4Aa has been determined by X-ray crystallography at a resolution of 2.8 Å (6). The structure of Cry4Aa is similar to the structures of previously characterized Cry toxins (24, 26, 31) that are composed of three domains (domains I, II, and III). In general, domain I, which is located in the N-terminal region, is composed of seven amphipathic α-helices and is thought to participate in membrane insertion. Domain II, which consists of three antiparallel β-sheets, is a putative receptor binding domain (Fig. (Fig.1).1). In particular, the loops in domain II that are exposed on the surface of the toxin molecule vary significantly in length and amino acid sequence among Cry toxins (31) and are thought to be receptor binding sites. Domain III in the C-terminal region contains two antiparallel β-sheets that form a β-sandwich fold with a jellyroll topology (31). Domain III is assumed to be involved in structural integrity, membrane protein recognition, or both (23, 24, 30).
The insecticidal mechanism of Cry toxin involves multiple steps, including ingestion by susceptible insects, solubilization in the alkaline midgut juice, activation by trypsin-like midgut proteases, binding to specific receptors on midgut epithelial cells, and then insertion into the plasma membrane followed by the formation of cation-selective channels or pores (26, 31, 34, 41). According to the colloid-osmotic lysis model, these channels or pores allow ions and water to pass into the cells, resulting in destruction of the membrane potential, cell swelling, cell lysis, and eventual death of the host (20, 21). Thus, the mechanism seems to be very complicated and is affected by multiple factors. The binding of the toxin to the specific receptor is considered a vital step for specific insecticidal activity (35). In fact, modification of the receptor molecules has been reported for insects resistant to certain Cry toxins (12, 22, 36).
In a search for the functional structures of Cry4Aa, we previously constructed various loop replacement mutants with mutations in the three major loops in domain II and showed that the replacement of loop 2 resulted in a significant loss of mosquitocidal activity. Replacement of loops 1 and 3 of Cry4Aa also affected mosquitocidal activity, but it did not eliminate it (17). In this study, to further characterize the loops, we constructed Cry4Aa mutants in which individual amino acids in the loops were replaced with alanine and analyzed the mutants to determine their mosquitocidal activity against Culex pipiens larvae. We also analyzed the structural integrity of the Cry4Aa mutant proteins subjected to proteolytic digestion and their binding affinity to brush border membrane vesicles (BBMV) prepared from C. pipiens larvae.
A synthetic cry4Aa (Cry4Aa-S1) gene (16), which showed hyperexpression in Escherichia coli, was used in this study. Mutations in loops 1, 2, and 3 were introduced by a 1-day mutagenesis procedure as reported previously (17, 18). The primers used for the mutagenesis are listed in Table Table1.1. Introduction of the mutations into loops 1, 2, and 3 was confirmed by DNA sequencing using an ABI PRISM 310 genetic analyzer (Applied Biosystems, Foster City, CA).
The Cry4Aa wild type and loop mutations were expressed as glutathione S-transferase (GST) fusions. E. coli BL21 cells harboring the corresponding mutant plasmids were cultured at 30°C in TB medium supplemented with ampicillin. Expression of the GST-Cry4Aa fusion was induced by adding isopropyl-β-d-thiogalactopyranoside (IPTG) at a concentration of 0.06 mM, followed by culturing for 3 h at 20°C. The E. coli cells were harvested by centrifugation and disrupted by sonication. The GST-Cry4Aa mutant proteins were purified using glutathione-Sepharose 4B (GE Healthcare Bio-Sciences AB, Uppsala, Sweden) as reported previously (16, 17, 39). The protein concentration was determined using a protein assay kit (Bio-Rad Laboratories, Inc., Hercules, CA) with bovine serum albumin as a standard. The GST-Cry4Aa mutant proteins were analyzed by 14% SDS-PAGE followed by visualization with Coomassie brilliant blue (CBB) staining.
The mosquitocidal activity of the GST-Cry4A mutant proteins was estimated using larvae of C. pipiens (fourth instar). Mosquito larvae were reared from eggs that were supplied by the Research and Development Laboratory, Dainihon Jochugiku Co., Ltd., Osaka, Japan. The bioassay used has been described previously (17). GST and wild-type GST-Cry4Aa were used as negative and positive controls, respectively. The concentration of the mutant GST-Cry4Aa proteins analyzed by SDS-PAGE was determined by densitometric scanning of the protein bands using a known amount of Cry4Aa as a standard. The mortality was recorded at 48 h after inoculation, and 50% lethal doses (LC50s) were determined using PROBIT analysis (14).
The structural stability of GST-Cry4Aa mutants subjected to trypsin treatment was analyzed. Since Cry toxins are activated by trypsin-like proteases in the midgut juice of susceptible insects, this assay is a presumptive test of folding fidelity (4). One microgram of purified mutant GST-Cry4Aa protein was treated with 20 ng of trypsin overnight at 37°C in 100 mM Tris-HCl (pH 8.0). The digests were analyzed by 14% SDS-PAGE followed by visualization using CBB staining.
BBMV were prepared from 7- to 10-day-old C. pipiens fourth-instar larvae as described previously (17, 32). BBMV containing 20 μg proteins were mixed with 2 μg of wild-type or loop mutant GST-Cry4Aa in phosphate-buffered saline containing protease inhibitor (Complete protease inhibitor cocktail tablets; Roche Applied Science, Mannheim, Germany) and then incubated for 1 h at room temperature. GST-Cry4Aa bound to BBMV was separated from unbound toxin by centrifugation and analyzed by SDS-PAGE as described previously (17).
Alanine substitution is widely used to analyze the biological functions of amino acid residues in protein molecules. Since alanine possesses a relatively small side chain, many target residues can be replaced with alanine without disrupting the overall structure of the protein molecule (7). In this study, amino acids in loops 1, 2, and 3 of Cry4Aa were individually replaced with alanine; alanine naturally occurs at A372 in loop 1 and at A511 in loop 3. As a result, 23 GST-Cry4Aa mutants (7 loop 1 mutants, 11 loop 2 mutants, and 5 loop 3 mutants) were constructed (Fig. (Fig.1).1). Most of these GST-Cry4Aa mutants were successfully expressed in E. coli. No difference in size or expression level was observed between wild-type and mutant GST-Cry4Aa (Fig. (Fig.2).2). However, the S428A loop 2 mutant did not accumulate protein. Western blotting using anti-Cry4Aa antiserum revealed that the S428A mutant protein was expressed but seemed to be degraded rapidly in E. coli cells (data not shown). The S428A mutant was therefore not used for further investigation. Since S428 is located in the vicinity of the β6 strand in domain II (Fig. (Fig.1),1), replacement of this residue may affect the integrity of the Cry4Aa structure, leading to rapid degradation.
E. coli cells expressing wild-type and mutant GST-Cry4Aa were bioassayed using C. pipiens larvae. The LC50s of the wild-type GST-Cry4Aa and mutant proteins were calculated based on the mortality of larvae at 48 h postinoculation. The LC50 of wild-type GST-Cry4Aa was determined to be 0.95 μg/ml, whereas GST alone showed no detectable toxicity.
Most loop 1 mutant proteins were as toxic as the wild type. The LC50s of the K371A, Q373A, T374A, T375A, P376A, and N378A mutants are shown in Table Table2.2. Only the N377A mutant showed a marginal decrease in toxicity (LC50, 2.88 μg/ml) compared to the wild type (Table (Table22).
On the other hand, the toxicity of the loop 2 mutants varied with the amino acid residue replaced. For example, the toxicities of the D430A mutant (LC50, 1.75 μg/ml), the K432A mutant (LC50, 1.59 μg/ml), the Y433A mutant (LC50, 1.12 μg/ml), and the D436A mutant (LC50, 0.91 μg/ml) were similar to that of the wild type. However, the other mutants (L429A, N431A, L434A, N435A, Y437A, and N438A) showed decreased toxicity (Table (Table2).2). Interestingly, many of the mutant proteins showing lower toxicity contained leucine and asparagine replacements. These residues in loop 2 may have some function in the mosquitocidal mechanism of Cry4Aa.
The toxicities of some loop 3 mutant proteins, including the I509A mutant (LC50, 1.21 μg/ml), the P510A mutant (LC50, 1.62 μg/ml), and the K514A mutant (LC50, 1.86 μg/ml), were similar to that of the wild type. However, the remaining mutant proteins (T512A and Y513A) showed decreased toxicity (Table (Table22).
GST-Cry4Aa mutant proteins were activated with trypsin and analyzed by SDS-PAGE. SDS-PAGE revealed active toxin fragments with molecular masses of 20 and 45 kDa for all mutants, as well as for wild-type GST-Cry4Aa (Fig. (Fig.3A).3A). This observation suggested that the structure of the GST-Cry4Aa mutant proteins was similar to that of wild-type GST-Cry4Aa. On the other hand, some unusual fragments with molecular masses ranging from 28 to 40 kDa were also observed in some mutants (Fig. (Fig.3A).3A). These unusual fragments may have resulted from the mutations introduced into the loops. However, the amounts of these fragments were so small that the mosquitocidal activity may not have been affected much.
We assayed the binding of loop 2 and 3 Cry4Aa mutants to BBMV prepared from C. pipiens larvae. The amount of Cry4Aa bound to BBMV varied according to the mutant (Fig. (Fig.3B).3B). However, no relationship was observed between the amount of Cry4Aa bound to BBMV and lethal activity against C. pipiens larvae. This finding was similar to our previous observations with loop replacement mutants (17).
We previously reported that replacement of loop 2 with loops 1 and 3 resulted in significant loss of mosquitocidal activity of Cry4Aa (17). However, in this study, we found that replacement of a single residue in loop 2 with alanine results in only a marginal decrease in Cry4Aa mosquitocidal activity. This suggested that in loop 2 of Cry4Aa some motif containing multiple amino acid residues rather than a single amino acid residue functionally determines the mosquitocidal activity. The involvement of loop 2 in determining the mosquitocidal activity was therefore analyzed further through construction of more mutants, in which multiple residues in loop 2 were replaced with alanine. Newly constructed mutants 4AL2Mm1 and 4AL2Mm2 had mutations in which six residues from S428 to Y433 and five residues from L434 to N438 were replaced with alanine, respectively. These mutants were successfully expressed as 90-kDa GST fusions in E. coli BL21 cells (Fig. (Fig.4A).4A). However, a bioassay using C. pipiens larvae revealed that the toxicity of these mutants was only marginally decrease (Fig. (Fig.4C).4C). The proteolytic digestion profile of the GST-Cry4Aa multiple mutants was also similar to that of wild-type GST-Cry4Aa (Fig. (Fig.4B4B).
Cry toxins have a narrow insecticidal spectrum, which suggests that there is a specific receptor for this family of toxins in susceptible insects. Therefore, binding to the specific receptor is an essential factor in determining the target insect spectrum and the toxicity of Cry toxins. In fact, a laboratory-selected Heliothis virescens strain was found to develop resistance to Cry1A due to alteration in the receptors for the toxin (22).
Involvement of the loops (loops 1, 2, and 3) in domain II in the toxin-receptor interaction has been reported for many Cry toxins (3, 25, 28, 29, 37). For example, in Cry3Aa, replacement of residues N353 and D354 in loop 1 with alanine results in loss of receptor binding ability, as well as toxicity against larvae of Tenebrio molitor and Leptinotarsa decemlineata (37). Similarly, replacement of W357 in loop 1 of Cry19Aa with alanine results in significant loss of toxicity against C. pipiens (29). The alanine substitution mutations Y410A, W416A, and D418A in loop 2 of Cry19Aa result in reduced toxicity against C. pipiens (>130-fold) and Aedes aegypti (4-fold) (29). Deletion of residues 365 to 371 in loop 2 of Cry1Aa or replacement of these residues with alanine eliminates nearly all toxicity against Bombyx mori (25). Mutants with the Q374N, P375A, W376Y, and P377A substitution mutations in Cry1C loop 2 exhibit reduced insecticidal activity against A. aegypti (3). Deletion of residues 440 to 443 in loop 3 results in a significant reduction in the toxicity of Cry1Ab against B. mori and Manduca sexta (28). Therefore, it can be inferred that the loop requirement for insecticidal activity varies with the individual combinations of Cry toxins and insect species. These observations also suggest that one of the three loops may be involved in receptor binding and toxicity.
In this study, we analyzed loops 1, 2, and 3 of Cry4Aa by using the alanine scanning method. Since replacement of Cry4Aa loop 2 with loop 1 or 3 resulted in a significant loss of mosquitocidal activity (17), we attempted to specify the residues in loop 2 that are essential for activity. Although a limited decrease in the mosquitocidal activity was observed for Cry4Aa mutants, all of the mutants still exhibited relatively high activity. Even the loop 2 multiple-replacement Cry4Aa mutants exhibited substantial levels of mosquitocidal activity. This suggests that the receptor binding site of Cry4Aa for C. pipiens larvae is different from loops 1, 2, and 3. Alternatively, multiple subsites that work cooperatively for receptor binding may be spread out in domain II and perhaps in domain III of Cry4Aa, as observed in Cry1Ac (10, 19) and Cry1C (9). Cry4Aa may thus differ from other well-characterized Cry toxins of B. thuringiensis in its receptor binding mechanism.
In this study, we also analyzed the binding of the loop 2 and 3 GST-Cry4Aa mutants to BBMV prepared from C. pipiens larvae (Fig. (Fig.3B).3B). However, no relationship between binding and toxicity was observed, suggesting that the binding assay data may have been obscured by nonspecific binding of GST-Cry4Aa. Nonspecific binding of Cry4Aa to BBMV of C. pipiens has been reported previously (40). If such binding occurs, we should rely on bioassay data rather than binding data to search the Cry4Aa molecule for the functional receptor binding site leading to mosquitocidal activity.
Many Cry toxins, including Cry4Aa, have very similar three-dimensional structures consisting of three domains. It is believed that this similarity in structure results in similar insecticidal mechanisms. Therefore, a functional structure, such as a receptor binding site, could be transferred from one Cry toxin to another. In fact, enhanced toxicity and a broadened insecticidal spectrum have been reported for loop-modified Cry toxins (1, 2, 38). Although the receptor binding site of Cry4Aa remains to be elucidated and the contradiction involving the necessity of loop 2 between our present and previous results (17) should be resolved, the fact that amino acid sequences in loops 1, 2, and 3 of Cry4Aa can be engineered without a loss of toxicity against C. pipiens suggests that these loops are an excellent target for modification to enhance the toxicity of Cry4Aa, as well as to broaden its insecticidal spectrum.
This work was supported in part by a Grant-in-Aid for Scientific Research (B) in Japan (grant 20380036).
Published ahead of print on 30 November 2009.