Bacillus thuringiensis Cry1Ab Domain III β-22 Mutants with Enhanced Toxicity to Spodoptera frugiperda (J. E. Smith)

Spodoptera frugiperda is an important worldwide pest of maize and rice crops that has evolved resistance to Cry1Fa-expressing maize in different countries. Therefore, identification of additional toxins with different modes of action is needed to provide alternative tools to control this insect pest. Bacillus thuringiensis (Bt) Cry1Ab and Cry1Ac toxins are highly active against several important lepidopteran pests but show varying and low levels of toxicity against different S. frugiperda populations. Thus, the identification of Cry1A mutants that gain toxicity to S. frugiperda and retain toxicity to other pests could be of great value to produce transgenic crops that resist a broader spectrum of lepidopteran pests. Here, we characterized Cry1Ab domain III β-22 mutants, and we found that a Cry1AbS587A mutant displayed increased toxicity against different S. frugiperda populations. Thus, Cry1AbS587A could be a good toxin candidate to produce transgenic maize with broader efficacy against this important insect pest in the field.

ABSTRACT

The fall armyworm, Spodoptera frugiperda, is an invasive maize pest that has spread from the Americas into Africa and Asia and causes severe crop damage worldwide. Most populations of S. frugiperda show low susceptibility to Bacillus thuringiensis (Bt) Cry1Ab or Cry1Ac toxins, which have been proved to be effective against several other lepidopteran pests. In addition, S. frugiperda has evolved resistance to transgenic maize expressing Cry1Fa toxin. The specificity and toxicity of Cry toxins are determined by their binding to different larval midgut proteins, such as aminopeptidase N (APN), alkaline phosphatase (ALP), and cadherin (CAD), among other proteins, by means of exposed domain II loop regions and also by the domain III β-sheets β-16 and β-22. Here, we analyzed different Cry1Ab mutants with mutations in the domain III β-22 region. Alanine-scanning mutagenesis of this region revealed that all mutants showed increased toxicity against a nonsusceptible Cry1Ab S. frugiperda population. Further analysis of the mutant toxin Cry1AbS587A (bearing a mutation of S to A at position 587) revealed that, compared to Cry1Ab, it showed significantly increased toxicity to three other S. frugiperda populations from Mexico but retained similar toxicity to Manduca sexta larvae. Cry1AbS587A bound to brush border membrane vesicles (BBMV), and its higher toxicity correlated with higher binding affinities to APN, ALP, and CAD recombinant proteins. Furthermore, silencing the expression of APN1 and CAD receptors in S. frugiperda larvae by RNA interference (RNAi) showed that Cry1AbS587A toxicity relied on CAD expression, in contrast to Cry1Ab. These data support the idea that the increased toxicity of Cry1AbS587A to S. frugiperda is in part due to an improved binding interaction with the CAD receptor.

IMPORTANCE Spodoptera frugiperda is an important worldwide pest of maize and rice crops that has evolved resistance to Cry1Fa-expressing maize in different countries. Therefore, identification of additional toxins with different modes of action is needed to provide alternative tools to control this insect pest. Bacillus thuringiensis (Bt) Cry1Ab and Cry1Ac toxins are highly active against several important lepidopteran pests but show varying and low levels of toxicity against different S. frugiperda populations. Thus, the identification of Cry1A mutants that gain toxicity to S. frugiperda and retain toxicity to other pests could be of great value to produce transgenic crops that resist a broader spectrum of lepidopteran pests. Here, we characterized Cry1Ab domain III β-22 mutants, and we found that a Cry1AbS587A mutant displayed increased toxicity against different S. frugiperda populations. Thus, Cry1AbS587A could be a good toxin candidate to produce transgenic maize with broader efficacy against this important insect pest in the field.

INTRODUCTION

Bacillus thuringiensis (Bt) is a Gram-positive bacterial species that produces diverse virulence factors for infecting different insect larval hosts (1, 2). Among these virulence factors, Cry toxins are pore-forming toxins that burst larval midgut cells by forming lytic pores in the plasma membrane (2, 3). Cry toxins have proven to be valuable tools, as spray formulations or expressed in genetically modified crops, for control of multiple agricultural pests or for killing mosquito vectors of different human diseases (3). Nevertheless, some crop pests show low susceptibility to Cry toxins (4).

Cry1A toxins are produced as protoxins of 130 kDa that are processed by midgut proteases of the susceptible larvae, leading to the formation of activated toxins of 65 kDa (1, 3). The activated toxins are composed of three domains. Domain I is a bundle of seven α-helices involved in toxin oligomerization, membrane insertion, and pore formation, while domains II and III, mainly composed of β-sheets, are involved in recognition of larval midgut proteins that trigger toxin oligomerization and membrane insertion necessary for the toxin pore formation activity (3). It was also shown that both Cry1A protoxins and activated toxins exert toxicity by independent pathways, and both forms of Cry1A proteins are able to bind to insect gut receptors, leading to different pore formation activities (5, 6). Domain II exposed loops are involved in binding to different larval midgut proteins, such as glycosyl-phosphatydil-inositol (GPI)-anchored proteins like alkaline phosphatase (ALP) and aminopeptidase N (APN), or to transmembrane proteins like cadherin (CAD) or ABCC2 transporter (2, 7–9), while domain III β-sheets β-16 and β-22 were shown to be involved in binding to ALP or APN (10–12).

Different strategies for improving Cry toxicity have been reported (13). These strategies include domain III shuffling among different toxins and mutagenesis of domain II exposed loops or domain III regions involved in receptor binding (13, 14). In the case of Cry1Ab and Cry1Fa toxins, it was shown that certain point mutations in domain III β-16, like Cry1AbN514A (bearing a change of N to A at position 514) or Cry1FaN507A, increased the toxicity of these mutants to different populations of Spodoptera frugiperda, a corn pest with worldwide importance (14). S. frugiperda is susceptible to Cry1Fa toxin but shows low and variable susceptibility to Cry1Ab or Cry1Ac toxins (4, 14, 15). It was reported that the high toxicity of Cry1AbN514A correlated with increased stability of this protein to the treatment with S. frugiperda midgut proteases, and this mutant also showed increased binding to brush border membrane vesicles (BBMV) due to improved binding to ALP, APN1, and especially CAD receptors (14).

Besides the Cry1Ab β-16 region of domain III, it was shown that the β-22 region is also involved in binding to ALP and APN in Bombyx mori and Manduca sexta larvae (10, 11). Here, we characterize several Cry1Ab mutants with mutations of different residues of the β-22 region from domain III and show that certain β-22 mutations, like Cry1AbS587A, also showed enhanced toxicity to different populations of S. frugiperda. The data show that its enhanced toxicity could be due in part to its enhanced binding affinity to the CAD midgut receptor protein of S. frugiperda.

RESULTS

Cry1Ab alanine substitutions in β-22 region increase toxicity to different S. frugiperda populations.

To determine the role of β-22 residues (583VFTLSAHV590) in toxicity to S. frugiperda, we performed alanine scanning of all β-22 residues except A588. Cry1Ab β-22 mutants were transformed into Bt for expression under sporulation conditions as described in Materials and Methods. The results in Fig. 1 show that Cry1AbV583A, Cry1AbT585A, Cry1AbS587A, Cry1AbH589A, and Cry1AbV590A produced a 130-kDa protoxin as the Bt strain transformed with Cry1Ab. Mutants Cry1AbF584A and Cry1AbL586A did not produce the 130-kDa protoxin and were not further characterized. To determine if any of the Cry1Ab β-22 mutants showed enhanced toxicity to S. frugiperda, as was shown previously for certain β-16 mutations (14), we performed bioassays against an S. frugiperda population (Sf-IBT) that has very low sensitivity to Cry1Ab (14). The data in Table 1 show that all of the Cry1Ab β-22 mutants, Cry1AbV583A, Cry1AbT585A, Cry1AbS587A, Cry1AbH589A, and Cry1AbV590A, showed enhanced toxicity to Sf-IBT, in contrast to Cry1Ab, which showed no toxicity to this population. Among all mutants tested, the Cry1AbS587A mutant showed the highest toxicity, as judged by the lack of overlapping values of the fiducial limits of the 50% lethal concentration (LC₅₀) value of this mutant (Table 1) with those of the rest of the mutants analyzed. This mutant was selected for further characterization, such as bioassays against M. sexta larvae, which are highly susceptible to Cry1Ab. The Cry1AbS587A mutant showed toxicity similar to that of Cry1Ab against M. sexta (LC₅₀ of 5.8 ng/cm² [confidence limits, 4.7 to 7.1] for the mutant toxin versus 3.4 ng/cm² [confidence limits, 2.6 to 4.3] for Cry1Ab). These data show that Cry1AbS587A retains toxicity similar to that of Cry1Ab to M. sexta.

FIG 1.

Production of Cry proteins from Cry1Ab domain III β-22 mutants. SDS-PAGE electrophoresis of spore/crystal suspensions from Cry1Ab and domain III mutants. First lane, molecular weight marker indicated in kDa; 2nd lane, Cry1Ab; 3rd lane, Cry1AbV583A; 4th lane, Cry1AbF584A; 5th lane, Cry1AbT585A; 6th lane, Cry1AbL586A; 7th lane, Cry1AbS587A; 8th lane, Cry1AbH589A; 9th lane, Cry1AbV590A.

TABLE 1.

Insecticidal activities of Cry1Ab and domain III mutant toxins against Sf-IBT population

Toxin	Mean LC50 in ng/cm2 (fiducial limits) for
	Spodoptera frugiperda (Sf-IBT)
Cry1Ab	>5,000
Cry1AbV583A	175 (106–331)
Cry1AbT585A	89 (64–122)
Cry1AbS587A	44 (26–63)
Cry1AbH589A	276 (160–619)
Cry1AbV590A	187 (135–265)

To determine if Cry1AbS587A showed enhanced toxicity to other S. frugiperda populations, we performed bioassays against four other S. frugiperda populations that were obtained from insects collected from different locations in Northern and Central Mexico. Some of these populations were previously characterized regarding to their susceptibility to Cry1Ab and Cry1AbN514A proteins (14). Here, we included a novel S. frugiperda population (Sf-UAEM1) raised from individuals collected in the state of Morelos, México, that was not reported before. In contrast to the Sf-IBT population, the other S. frugiperda populations were susceptible to Cry1Ab at different levels (Table 2). The results in Table 2 show that the susceptibilities of the different S. frugiperda populations ranged from 131 to 1,500 ng/cm² and the Sf-UAEM1 population showed the highest susceptibility to Cry1Ab. In addition to the Sf-IBT population, the Cry1AbS587A mutant showed significantly increased toxicity against three of the other populations analyzed, since it increased the toxicity against Sf-Valle del Fuerte 5.9-fold, against the SfS-Mex population 9.4-fold, and against the Sf-UAEM1 population 3.8-fold. The only population that showed reduced toxicity (3-fold) to Cry1AbS587A compared to that of Cry1Ab was the Sf-La Laguna population (Table 2). Previously, we showed that the domain III β-16 Cry1AbN514A mutant also showed similar enhanced toxicities to the Sf-Valle del Fuerte and SfS-Mex populations (14). Thus, we constructed a double Cry1AbN514-S587A mutant to determine if the double mutation could increase the toxicity of Cry1Ab to these S. frugiperda populations even more. The results in Table 2 show that Cry1AbN514A-S587A showed 2-fold-higher LC₅₀ values to Sf-Valle del Fuerte, SfS-Mex, and Sf-UAEM1 but that these values were not significantly different, since their fiducial limits overlapped. Regarding Sf-La Laguna, Cry1AbN514A-S587A showed toxicity similar to that of Cry1Ab. These data show that the double mutant Cry1AbN514A-S587A did not improve the toxicity of the single mutants Cry1AbN514A and Cry1AbS587A against S. frugiperda (Table 2) (14).

TABLE 2.

Toxicity of Cry1Ab and domain III mutants against different S. frugiperda populations from Mexico

Population	Mean LC50 (ng/cm2) (fiducial limits)
	Cry1Ab	Cry1AbS587A	Cry1AbN514A-S587A
Sf-Valle del Fuerte	1,420 (752–3,730)a	240 (179–330)	660 (224–4,289)
Sf-La Laguna	482 (326–802)a	1,481 (693–4,748)	411 (213–869)
SfS-Mex	1,590 (744–5,480)a	169 (86–457)	476 (301–846)
Sf-UAEM1	131 (93–212)	34 (25–43)	75 (58–100)

^aData obtained at the same time as that for Cry1AbS587A and Cry1AbN514A-S587A but published previously (14).

Enhanced toxicity of Cry1AbS587A correlated with increased binding to S. frugiperda BBMV and increased binding affinity to APN1, ALP, and CAD receptors.

To determine the effect of the Cry1AbS587A mutation on receptor binding, we performed qualitative binding assays to BBMV isolated from Sf-IBT of Cry1Ab and of the other three β-22 mutants that also showed higher toxicity against this S. frugiperda population (Table 1). The results in Fig. 2 show that Cry1Ab did not bind to BBMV of this population, in contrast to Cry1AbV583A, Cry1AbT585A, or Cry1AbS587, which were able to bind to BBMV from Sf-IBT. In addition, we further determined the binding of the Cry1AbS587A mutant to the full-length APN1 or ALP or to a fragment of CAD that contains the Cry1Ab binding sites, using enzyme-linked immunosorbent assays (ELISAs). These proteins were previously cloned from the Sf-IBT population into Escherichia coli to produce them as recombinant proteins (14). The results in Fig. 3 show that Cry1Ab bound APN1, ALP, and CAD with relatively low affinities (APN1 dissociation constant [K_d] = 147 nM, ALP K_d = 90 nM, and CAD K_d = 147 nM). In contrast, the Cry1AbS587A mutant showed 5-fold higher binding affinities to these receptors (APN1 K_d = 29 nM, ALP K_d = 19 nM, and CAD K_d = 29 nM) (Fig. 3). These results suggest that the binding of Cry1AbS587A to BBMV compared to that of Cry1Ab observed in the experiment whose results are shown in Fig. 2 may be due to the enhanced binding affinity of this mutant to APN1, ALP, and CAD midgut receptors.

FIG 2.

Binding of biotin-labeled Cry1Ab and β-22 mutants to BBMV from S. frugiperda population Sf-IBT. Ten micrograms of BBMV protein were incubated with 5 nM biotin-labeled Cry1Ab activated proteins, and bound toxin was recovered by centrifugation. The toxin bound to BBMV was separated by electrophoresis in 10% SDS-PAGE, electrotransferred to PVDF membranes, and revealed with streptavidin-horseradish peroxidase as described in Materials and Methods. First lane, Cry1Ab; 2nd lane, Cry1AbV583A; 3rd lane, Cry1AbT585A; 4th lane, Cry1AbS587A.

FIG 3.

Binding of Cry1Ab and Cry1AbS587 to recombinant APN1, ALP, and Cad receptors. The recombinant protein receptors were fixed (2.5 μg) to ELISA plates, incubated with different concentrations of Cry1Ab or Cry1AbS587A, and revealed with anti-Cry1Ab antibody coupled to a secondary antibody with horseradish peroxidase. The results are the mean values ± standard deviations (SD) from three replicates. Binding assays of Cry1Ab and Cry1AbS587A were performed at the same time, but the Cry1Ab data were published previously (14).

Cry1AbS587A relies on CAD receptor for toxicity to S. frugiperda, in contrast to Cry1Ab.

In order to determine the role of CAD and APN1 in the toxicity of Cry1AbS587A to S. frugiperda, we performed gene-silencing experiments of CAD and APN1 expression by feeding larvae cad double-stranded RNA (dsRNA) or apn1 dsRNA, as previously reported (16), and compared the effects of silencing the expression of these proteins on Cry1Ab and Cry1AbS587A toxicity. In addition, we also analyzed the effect of silencing both receptors on the toxicity of Cry1AbN514A, which also showed enhanced toxicity to the Sf-UAEM1 population, with an LC₅₀ of 59 ng/cm² (fiducial limits, 43 to 81), and was previously shown to have enhanced toxicity to the other S. frugiperda populations (14). We decided to perform the silencing experiments in the Sf-UAEM1 population, since this population is more susceptible to Cry1Ab than Sf-IBT, which showed no Cry1Ab susceptibility. The Sf-UAEM1 population showed a significant, 4-fold-higher susceptibility to Cry1AbS587A than to the Cry1Ab toxin (Table 2). Our hypothesis was that silencing these receptors might affect the toxicity of the mutant toxins (Cry1AbS587A and Cry1AbN514A), since both of them showed increased binding affinity to these receptors. For these assays, we elected to use a single dose in the toxicity assays. The selected dose is close to 50% mortality for the wild-type Cry1Ab toxin, and thus, the possible role of the silenced protein cold be clearly determined as a decrease or an increase in the toxicity after silencing. The results in Fig. 4A show a quantitative PCR (qPCR) analysis of apn1 and cad transcripts after silencing their expression, showing that both cad and apn1 gene expression were significantly lower in the insects treated with their corresponding dsRNA, in contrast to their expression in control insects that were not treated with any dsRNA (P < 0.001). The results in Fig. 4B show that at the selected concentration (2.5 μg/cm²), Cry1Ab killed 31% of the control larvae with no dsRNA treatment; these control larvae were fed the E. coli lysate with empty vector as described in Materials and Methods. In contrast, the Cry1AbN514A and Cry1AbS587A mutants resulted in 92% or 89% mortality, respectively, in larvae exposed to the same 2.5-μg/cm² dose, which is in agreement with their higher toxicity against the Sf-UAEM1 population (Table 2). However, CAD-silencing experiments showed that Cry1Ab toxicity was slightly reduced when the CAD expression was silenced, since larvae treated with cad dsRNA showed 21% mortality, compared to 31% mortality of the control larvae. However, statistical analyses indicated that these differences in the toxicity of Cry1Ab to the CAD-silenced larvae and control (nonsilenced) larvae were not statistically significant (P > 0.1).

FIG 4.

Silencing of APN1 and CAD receptors shows that Cry1Ab domain III mutants rely on the CAD receptor for toxicity. (A) apn1 or cad transcript abundance was determined by qPCR and SYBR green as described in Materials and Methods. The transcript levels of the ribosomal protein S3 gene (rps) were used to normalize gene transcript levels. The control was the transcript level of either apn1 or cad normalized to 1 in larvae fed with normal diet. Data are mean transcript levels of 12 independent midguts from control larvae or larvae treated with dsRNA. Different numbers of asterisks indicate significant differences (P < 0.001). (B) Mortality assays of larvae treated with no dsRNA (Control), APN1 dsRNA, and cad dsRNA. Larvae (n = 72) were exposed to 2.5 μg of each toxin, and mortality was scored after 7 days as described in Materials and Methods. Different letters denote significantly different values. Data are mean values ± SD. P values were calculated using Student’s t test for two groups, control larvae versus treated larvae (dsAPN1 or dsCAD) (https://www.graphpad.com/quickcalcs/ttest1/), and P < 0.05 was considered significantly different.

In contrast, Cry1AbN514A and Cry1AbS587A mutant proteins both relied largely on CAD expression, since cad dsRNA-treated larvae showed only 35 to 40% mortality for both domain III mutants compared to 90% mortality in control larvae, suggesting a 60% higher tolerance of CAD-silenced larvae than of the control larvae to Cry1AbN514A or Cry1AbS587A proteins. Statistical analyses indicated that the difference in toxicity of both Cry1Ab domain III mutants (Cry1AbN514A and Cry1AbS587A) against CAD-silenced larvae than against control (nonsilenced) larvae is extremely statistically significant (P < 0.0012).

Regarding APN1 silencing, larvae fed apn1 dsRNA showed similar susceptibilities to 2.5 μg/cm² Cry1Ab (31% mortality in control larvae and 41% mortality in the APN1-silenced larvae; P < 0.04). Likewise, similar susceptibilities were also observed after treatment with the Cry1AbN514A or the Cry1AbS587A mutant against the APN1-silenced versus control larvae (P > 0.17) (Fig. 4B).

DISCUSSION

S. frugiperda is an important crop pest worldwide that has evolved resistance to Cry1Fa toxin-expressing maize in Puerto Rico, the United States, Brazil, and Argentina (17–20). This pest is an invasive insect that recently has migrated to Africa and Asia, jeopardizing the production of maize and other crops in multiple countries, since it is a polyphagous pest (21, 22). Therefore, additional toxins with different modes of action are needed to provide alternative tools to control this insect pest. Previous work revealed that certain Cry1Ab β-16 mutations (for example, Cry1AbN514A) showed increased toxicity to several S. frugiperda populations and retained toxicity to M. sexta (14). Since domain III regions of Cry1Ab toxin, such as the β-16 and β-22 regions, have been shown to be involved in receptor binding, we performed β-22 alanine-scanning mutagenesis to search for Cry1Ab mutants with enhanced toxicity to S. frugiperda. For screening purposes, we used an S. frugiperda population that was previously shown to have low susceptibility to Cry1Ab (Sf-IBT) (14, 23). We showed that mutants with alanine substituted for each of the β-22 residues gained toxicity to Sf-IBT (Table 2). However, Cry1AbS587A showed the highest toxicity to Sf-IBT without losing toxicity to M. sexta and was further characterized.

Cry1AbS587A showed enhanced toxicity to three other S. frugiperda populations obtained from different regions of Mexico (14). Interestingly, it showed 3-fold-lower toxicity to the Sf-La Laguna population, which also showed reduced susceptibility to the β-16 Cry1AbN514A mutant, as previously reported (14). The reason for the different susceptibilities to Cry1Ab domain III mutations in the Sf-La Laguna population compared to those of the other populations analyzed still remains to be determined. We were unable to analyze the binding of Cry1Ab, Cry1AbN514A, and Cry1AbS587A to BBMV from the Sf-La Laguna population to determine if the different susceptibilities of this population to these proteins is related to effects in toxin binding because this population is not maintained anymore in the laboratory. Also, it is important to know the frequency of other populations with similar characteristics to Sf-La Laguna, since if this population is more widespread, the usefulness of Cry1Ab mutants in the control of S. frugiperda will be reduced. It has been reported that different S. frugiperda populations from diverse Latin American countries or from the same country may show very different susceptibilities to Cry1 toxins or to chemical insecticides, indicating a great genetic variability among S. frugiperda populations (24–27).

The double Cry1AbN514A-S587A domain III mutant did not further improve Cry1Ab toxicity to S. frugiperda populations (Table 2). Interestingly, the double mutant showed toxicity similar to that of the Cry1Ab toxin. It has been shown that in the 3-dimensional structure of Cry1Aa, these two β-strands (β-16 and β-22) of domain III are localized close together in the same face of the toxin, suggesting that both regions may be involved in the same binding interactions (10). Our data indicate that increased toxicity of Cry1AbS587A to S. frugiperda correlated with its increased capacity to bind to BBMV from the Sf-IBT population compared to Cry1Ab, which did not bind to BBMV from this population (Fig. 2). However, it is important to mention that in these binding assays, we analyzed total binding, which may include nonspecific binding to the membranes that could also be in part responsible for the increased binding of Cry1AbS587A. To determine if increased binding of Cry1AbS587A or Cry1AbN514A is responsible for their increased toxicity against the other S. frugiperda populations, it will be important in the future to also analyze specific binding of these toxins to BBMV by performing competition binding analysis, since the other populations analyzed were shown to be susceptible to Cry1Ab. In addition, we observed increased binding affinity to APN1, ALP, and CAD receptors, similar to that previously reported for Cry1AbN514A (14). It is important to mention that these receptor proteins were not purified from insect larvae; in contrast, they were expressed in E. coli cells. However, we have previously shown that these Cry receptors produced in E. coli cells are able to bind Cry1Ab toxin in a saturable way with binding affinities in the nanomolar range, which agrees with the binding affinities for similar proteins isolated from lepidopteran insects like M. sexta, suggesting that E. coli-produced Cry receptor proteins have no major conformational issues (12).

Gene silencing experiments showed that the increased toxicity of Cry1AbN514A or Cry1AbS587A to the Sf-UAEM1 population is likely due to CAD binding, since no effect on toxicity of these two mutants was observed in APN1-silenced larvae, in contrast to CAD-silenced larvae, in which the toxicity of both Cry1AbN514A and Cry1AbS587A mutant proteins was significantly reduced (Fig. 4). In addition, Cry1Ab toxicity was not significantly changed in CAD-silenced larvae (Fig. 4). The lack of effect of CAD silencing on Cry1Ab toxicity is in agreement with a recent report that shows that S. frugiperda CAD is not a functional receptor for Cry1Ab or Cry1Fa toxins, according to its analysis of an S. frugiperda CAD knockout mutant that was genome edited using clustered regularly interspaced short palindromic repeat (CRISPR)-CRISPR-associated protein 9 (Cas9) (28). Overall, the data presented here suggest that increased binding affinity of Cry1AbS587A to CAD is responsible in part for its higher toxicity to Sf-UAEM1. However, there was still considerable toxicity (35 to 40% mortality) of the Cry1AbN514A and Cry1AbS587A toxins in CAD-silenced larvae, suggesting that other receptors besides CAD also contribute to the enhanced toxicity of these toxins. It has been shown that ABCC2 is linked to resistance to Cry1Fa toxin in S. frugiperda, and the expression of S. frugiperda ABCC2 in SF9 cells conferred susceptibility to Cry1Fa and also to Cry1Ab and Cry1Ac toxins (29). Thus, it is still possible that the enhanced toxicity of Cry1AbN514A or Cry1AbS587A could also involve enhanced binding to ABCC2. This remains to be determined.

Regarding APN1, although Cry1AbS587A gained binding affinity to APN1 and to ALP, silencing APN1 had no effect on toxicity (Fig. 4). This could be explained since it has been suggested that APN and ALP binding have redundant roles in the Cry1Ab mode of action (23), which could explain the lack of phenotype in the APN1-silenced larvae. Cry1Ab β-16 and β-22 have been shown to be involved in APN and ALP binding (10, 11). The increased binding affinity of Cry1AbS587A to CAD could suggest that domain III β-22 might also be involved in CAD binding. However, previously reported experimental evidence suggests that Cry1Ab binds CAD by domain II loop 2 and loop 3 exposed regions and that β-16 and β-22 are not involved in binding to this receptor (7–9). As previously discussed for the Cry1AbN514A mutation (14), it is possible that subtle structural changes caused by domain III mutations have consequences for the conformation of domain II loop regions, favoring binding to CAD receptor. Whether domain III is directly involved in CAD binding or domain III mutations have subtle consequences for the domain II loop structure remains to be analyzed. In any case, Cry1AbS587A could be a good toxin candidate to produce transgenic maize with broader efficacy against S. frugiperda pests in the field. We show here that this mutant has improved toxicity against different S. frugiperda populations. However, more S. frugiperda populations, including those resistant to Cry1Fa, should be tested in the future to determine the usefulness of Cry1Ab domain III mutants. Finally, it still remains to be determined if Cry1AbS587A shares receptors with Cry1Fa or with Cry1A.105 to determine its potential use in pyramided crops.

MATERIALS AND METHODS

Insect populations.

All S. frugiperda populations except Sf-UAEM1 were described previously (14, 30). The Sf-IBT population was traditionally maintained at the Instituto de Biotecnología, UNAM (IBT-UNAM) until December 2018 when this population collapsed. The SfS-Mex, Sf-La Laguna, and Sf-Valle del Fuerte populations were maintained at Centro Universitario UAEMex. The Sf-La Laguna population is also no longer maintained in the laboratory. The Sf-UAEM1 population was obtained from individuals obtained in 2017 from maize fields from Puente Ixtla in the state of Morelos, México, and maintained at Centro de Investigaciones Biológicas, UAEM. The Manduca sexta Ms-IBT population was maintained at IBT-UNAM without exposure to Cry toxins. The growth conditions for all insect populations were 25°C with relative humidity of 75% and a light-dark photoperiod of 13:11 h.

Site-directed mutagenesis.

For mutagenesis of the cry1Ab gene, plasmid pHT315-Cry1Ab (31) was used as the template. The QuikChange multisite-directed mutagenesis kit was used, following the manufacturer’s instructions (Stratagene). Specific oligonucleotides were designed for alanine replacement of all β-22 residues (Table 3). Plasmids with the desired mutations were identified after DNA sequencing, performed at the DNA sequencing facilities of IBT-UNAM. For Bacillus thuringiensis (Bt) transformation, plasmids were first transformed into E. coli SCS110 (dam dcm mutant) for purification of plasmids lacking methylation, which is required for efficient Bt transformation. Transformation into the Bt 407 cry mutant strain (32) was done by electroporation as reported previously (33).

TABLE 3.

Oligonucleotides used in this study

Primer	Sequence
V583A	5′-TCAAATGGATCAAGTGCATTTACGTTAAGTGCT-3′
F584A	5′-AATGGATCAAGTGTAGCTACGTTAAGTGCTCAT-3′
T585A	5′-GGATCAAGTGTATTTGCGTTAAGTGCTCATGTC-3′
L586A	5′-TCAAGTGTATTTACGGCAAGTGCTCATGTCTTC-3′
S587A	5′-AGTGTATTTACGTTAGCGGCTCATGTCTTCAAT-3′
H589A	5′-TTTACGTTAAGTGCTGCTGTCTTCAATTCAGGC-3′
V590A	5′-ACGTTAAGTGCTCATGCATTCAATTCAGGCAAT-3′
CadSf28iRev	5′-TCAGGAATTCCGAAAATTAACTACCAAGGAATCA-3′
CadSf28iFor	5′-AACGAAGCTTGATAACTCATAACACTTGTACTGAT-3′
CadSfqRev	5′-GCTGGATATCGTGAACATCG-3′
CadSfqFor	5′-CGTCTCCTCTCTCGCTGTCT-3′

Production of Cry1Ab and Cry1Ab β-22 mutants.

All Bt 407 strains containing pHT315-Cry1Ab or mutant plasmids were grown in nutrient sporulation medium (34) plates supplemented with erythromycin (10 μg/ml) and incubated at 30°C until sporulation was completed (usually 3 days). Spore/crystal suspensions were washed three times in 0.3 M NaCl, 10 mM EDTA, pH 8.0, and then three times with 1 mM phenylmethylsulfonyl fluoride (PMSF) and stored at 4°C. The protein concentration was determined by Bradford assays using bovine serum albumin (BSA) as a standard. The soluble protoxins were obtained by the suspension of spores/crystals in 50 mM Na₂CO₃/NaHCO₃ buffer containing 0.02% mercaptoethanol, pH 10.5, and incubated at 37°C for 2 h. After centrifugation at 16,873 × g, the insoluble pellets were discarded. The soluble protoxins were proteolytically activated with trypsin (Sigma-Aldrich Co., St. Louis, MO) (1:50 [wt/wt] trypsin/Cry1Ab) at 37°C for 2 h. The reaction was stopped with 1 mM PMSF. Activated samples were analyzed by SDS-PAGE, and protein concentration was determined by the Bradford assay.

BBMV.

Midgut tissue from third-instar larvae of S. frugiperda (Sf-IBT) were dissected. Brush border membrane vesicles (BBMV) were prepared by a previously described differential precipitation procedure (35). BBMV were stored at −70°C.

Qualitative binding assays.

For binding assays, activated Cry1Ab or Cry1Ab mutants were labeled with biotin-N-hydroxysuccinimide ester (Amersham Biosciences, Waltham, MA), following the manufacturer’s procedure. Cry1Ab or Cry1Ab mutant proteins labeled with biotin (5 nM) were incubated with BBMV (10 μg protein) in binding buffer (phosphate-buffered saline [PBS], 0.1% BSA, 0.1% Tween 20) for 1 h at room temperature as previously described (14). After binding, samples were centrifuged for 10 min at 117,000 × g and the pellet was boiled for 3 min in loading buffer before separation by 10% SDS-PAGE. After electrophoresis, gels were transferred to polyvinylidene difluoride (PVDF) membrane (Millipore, Burlington, MA) as described previously (14). Blotted membranes were blocked with 2% Tween 20/PBS solution for 20 min and washed twice with 0.1% Tween 20/PBS. Biotin-labeled proteins were revealed by horseradish peroxidase (HRP)-streptavidin (Millipore, Burlington, MA) (1:5,000 dilution in 0.1% Tween 20/PBS), incubated for 1 h at room temperature, and then washed twice and then two times with PBS. The membrane was revealed with ImmunoCruz Western blotting luminol (Santa Cruz Biotechnology, Dallas, TX).

Production of recombinant proteins.

Recombinant APN1 (accession number MT673677) or CAD fragment CR7-12 (accession number AX147205.1) cloned in pET22b plasmid or ALP (accession number MT506049) cloned in pET28a was produced in E. coli strain BL21(DE3) as previously reported (14). Briefly, 100 μl of overnight culture of LB broth with 50 μg/ml ampicillin in the case of APN1 or CAD or 50 μg/ml kanamycin in the case of ALP was used to inoculate 100 ml of 2× TY broth medium (8 g Bacto tryptone, 5 g yeast extract, 2.5 g/liter NaCl, pH 7) in a 250-ml flask with either 100 μg/ml ampicillin for APN1 or CAD or 100 μg/ml kanamycin for ALP and incubated at 37°C with constant agitation (200 rpm) until an optical density at 600 nm (OD₆₀₀) of 0.6 was reached (usually 2 h). Expression of recombinant proteins was induced by the addition of 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside; Sigma), and the mixture was incubated with agitation for 5 h at 30°C. After centrifugation of cell cultures at 16,873 × g for 15 min at 4°C, the cell pellet was suspended in 10 mM Tris-HCl, 1 mM EDTA, 8 M urea, pH 8, and sonicated for 5 min on ice. The cell debris was separated by centrifugation at 117,000 × g for 30 min at 15°C, and the supernatant was passed through a Ni-agarose bead column. PBS, pH 7.5, with 35 mM imidazole was used for washing the column, and finally, the recombinant proteins were eluted by using 250 mM imidazole in PBS, pH 7.5. The eluted proteins were dialyzed with PBS and analyzed by SDS-PAGE using 10% acrylamide gels.

ELISA binding assays.

ELISA plates (Nunc) were coated with 1 μg of APN1, ALP, or CAD recombinant protein overnight at 4°C. Unbound toxin was removed by washing with PBS and blocked with blocking buffer (2% skim milk in 1× PBS) for 2 h at 37°C. Different concentrations of Cry1Ab or Cry1AbS587A were incubated with the coated ELISA plates, and unbound toxin was removed by washing three times with PBS. The bound toxins were revealed by using anti-Cry1Ab antibody (1:20,000) and anti-rabbit secondary antibody conjugated with HRP (Sigma). Finally, the plates were revealed with 0.5 mg/ml o-phenylenediamine (Sigma) and 0.0075% H₂O₂ followed by the addition of 50 μl of 1 M H₂SO₄, and the OD₄₉₀ was determined using an ELISA microplate reader. All experiments were performed in triplicate, and relative binding affinities were calculated using Scatchard analysis with SigmaPlot (Systat Software, Inc.).

Gene silencing in S. frugiperda.

For gene silencing, a 378-bp fragment of apn1 was cloned into the pLITMUS28i vector, which contains two T7 promoters in an inverted orientation (New England Biolabs), as previously described (16). For CAD silencing, primers with EcoRI and HindIII restriction sites were designed (Table 3) to amplify a 369-bp fragment by PCR, and the PCR product was also cloned into the pLITMUS28i vector. Plasmid pLITMUS28i:apn1 or pLITMUS28i:cad was transformed into E. coli strain HT115 for dsRNA production as previously described (16). One hundred microliters of overnight culture of HTT115/pLITMUS28i:apn1 or HTT115/pLITMUS28i:cad grown in LB broth supplemented with 100 μg/ml ampicillin was used to inoculate 50 ml of the same medium and grown at 37°C until an OD₆₀₀ of 0.6 was reached. dsRNA synthesis was induced in these bacteria by adding 0.1 mM IPTG, and the cell culture was incubated for an additional 4 h at 30°C. For dsRNA quantification, the RNeasy minikit (Qiagen) was used for dsRNA extraction from 1 ml of the cell culture, following the manufacturer’s instructions. The RNA samples were separated on agarose gels, and dsRNA concentrations were estimated by comparison with an RNA molecular-weight marker. The rest of the cell culture was centrifuged at 16,873 × g, and the cell pellet was suspended in PBS, pH 7.4 (5 ml). The suspended cells were sonicated 10 times at 95% intensity for 1 min, with a resting time on ice of 2 min between each sonication step. For delivering the dsRNA to the larvae, 35 μl of the sonicated sample was applied per well over the surface of artificial diet in 24-well plates (Corning Glass Works, Corning, NY), and one neonate larva was added per well. For bioassays, the treated larvae and controls treated with normal diet without dsRNA were separated into two groups. One group of 24 larvae was placed in wells containing diet with 2.5 μg/cm² of Cry1Ab or Cry1AbS587A spores/crystals, while the other group of 24 larvae was placed on diet without toxin. Mortality was recorded after 7 days. At least three replicates of these bioassays were performed, and data are shown as mean values ± standard deviations (Fig. 4B). The P value was calculated using Student’s t test for two groups, control versus treated larvae (dsAPN1 or dsCAD) (https://www.graphpad.com/quickcalcs/ttest1/), and P < 0.05 was considered significantly different.

qPCR analysis.

From the group of silenced larvae that were not exposed to toxins, midgut tissue was dissected from 12 treated larvae or 12 controls and samples from each midgut were suspended in RNAlater. Then, total RNA was extracted using RNeasy (Qiagen). The qPCR assays were performed in the Eco real-time PCR system (Illumina) using SYBR green PCR master mix (ThermoFisher Scientific) with the oligonucleotides listed in Table 3. PCR conditions were 95°C for 10 min and then 40 cycles at 95°C for 30 s, 60°C for 60 s, and 95°C for 15 s, with a final primer concentration of 250 nM. All qPCRs were performed in triplicate for each gene of each experimental group. The mean silencing was calculated from the 12 midgut samples, and standard deviations were calculated. Three biological replicates were done. Relative expression levels were normalized using the rps3 ribosomal gene, and the data were analyzed by the cycle threshold (2^−ΔΔCT) method as reported previously (16).

Mortality bioassays.

Toxicity assays were performed as previously reported (14). To determine the toxicity of Cry1Ab and β-22 mutants, five doses of spores/crystals from 0.1 to 1,000 ng/cm² were applied to insect diet using polystyrene plates with 24 wells (Cell Wells; Corning Glass Works, Corning, NY). Plates containing 24 larvae exposed to the same dose of toxin were assayed in triplicate, making 72 larvae per dose of toxin. All bioassays were performed at 28°C with 65% ± 5% relative humidity and a light-dark photoperiod of 16:8 h. Bioassays performed at Centro de Investigaciones Biológicas, UAEM, were performed using 128-well polystyrene plates (Bio-Assay tray Bio-BA-128; C-D International, Inc.) using similar doses of toxins as described above. Four replicates were performed for each dose of toxin, and mortality was recorded after 7 days. LC₅₀ values and fiducial limits were calculated using Probit software (Polo; LeOra).