Critical Domains in the Specific Binding of Radiolabeled Vip3Af Insecticidal Protein to Brush Border Membrane Vesicles from Spodoptera spp. and Cultured Insect Cells

ABSTRACT

Vegetative insecticidal proteins (Vip3) from Bacillus thuringiensis have been used, in combination with Cry proteins, to better control insect pests and as a strategy to delay the evolution of resistance to Cry proteins in Bt crops (crops protected from insect attack by the expression of proteins from B. thuringiensis). In this study, we have set up the conditions to analyze the specific binding of ¹²⁵I-Vip3Af to Spodoptera frugiperda and Spodoptera exigua brush border membrane vesicles (BBMV). Heterologous competition binding experiments revealed that Vip3Aa shares the same binding sites with Vip3Af, but Vip3Ca does not recognize all of them. As expected, Cry1Ac and Cry1F did not compete for Vip3Af binding sites. By trypsin treatment of selected alanine mutants, we were able to generate truncated versions of Vip3Af. Their use as competitors with ¹²⁵I-Vip3Af indicated that only those molecules containing domains I to III (DI-III and DI-IV) were able to compete with the trypsin-activated Vip3Af protein for binding and that molecules only containing either domain IV or domains IV and V (DIV and DIV-V) were unable to compete with Vip3Af. These results were further confirmed with competition binding experiments using ¹²⁵I-DI-III. In addition, the truncated protein ¹²⁵I-DI-III also bound specifically to Sf21 cells. Cell viability assays showed that the truncated proteins DI-III and DI-IV were as toxic to Sf21 cells as the activated Vip3Af, suggesting that domains IV and V are not necessary for the toxicity to Sf21 cells, in contrast to their requirement in vivo.

IMPORTANCE This study shows that Vip3Af binding sites are fully shared with Vip3Aa, only partially shared with Vip3Ca, and not shared with Cry1Ac and Cry1F in two Spodoptera spp. Truncated versions of Vip3Af revealed that only domains I to III were necessary for the specific binding, most likely because they can form the functional tetrameric oligomer and because domain III is supposed to contain the binding epitopes. In contrast to results obtained in vivo (bioassays against larvae), domains IV and V are not necessary for ex vivo toxicity to Sf21 cells.

INTRODUCTION

Globally, an estimated 20 to 40% of crops are lost due to pests and diseases (1), with insects representing a significant portion of this loss, both by direct damage and indirectly through the transmission of plant diseases (2). To date, the Gram-positive bacterium Bacillus thuringiensis (Bt) is known as the most economically successful entomopathogen, and it has been estimated to account for the 75 to 95% of the microbial biopesticide market (3). Also, Bt insecticidal genes encoding different Cry proteins have been expressed in plants (Bt crops) since 1996, and the area planted with Bt crops represents >53% of the global cultivated area of genetically modified crops in the world (4–6). More recently, vip3 genes have been introduced in elite crop varieties to complement the insecticidal action of the Cry proteins, as well as to help manage the development of insect resistance (7, 8).

Vip3 proteins contain from 787 to 789 amino acids, with a molecular mass of 89 kDa. Three main groups have been described: Vip3A, Vip3B, and Vip3C (9). The mode of action of Vip3 proteins has been studied mainly with Vip3Aa protein. It is well accepted that Vip3 proteins are produced as full-length proteins (protoxins), and after ingestion, they are activated by midgut proteases, which cleave the N-terminal 199 amino acids from the rest of the protein, generating two fragments (19 and 65 kDa) that remain attached to exert their toxic activity. This has been shown by different means (10–12) and more recently by the three-dimensional (3D) structure of the activated Vip3Aa (13). The activated protein crosses the peritrophic matrix and binds to the apical membrane of brush border epithelial cells, inducing pore formation, which eventually leads to the death of the larvae (14–16). Several insect proteins have been identified to serve as putative receptors of Vip3Aa: the ribosomal protein S2 from Sf21 cells (17), the scavenger receptor class C protein and fibroblast growth factor receptor-like proteins from cultured Sf9 cells (18, 19), and the 48-kDa tenascin protein from Agrotis ipsilon brush border membrane vesicles (BBMV) (20). In addition to pore formation, induction of apoptosis and internalization of the bound Vip3Aa have been observed (21–24).

The 3D structure of the Vip3Aa and Vip3B proteins shows that they have a tetrameric organization (13, 25), with each monomer containing five structural (and most likely, functional) domains (DI to DV) (13, 25–29). The highly conserved N-terminal domains (DI to DIII) have been shown to play a crucial role in maintaining the integrity of the tetramer (29), whereas the two most C-terminal domains (DIV to DV), which are highly variable, do not seem to be necessary for maintaining the tetrameric structure (13, 29), though they are critical for the insecticidal activity in vivo (26, 29–32). Domains IV and V are carbohydrate-binding modules (CBM), and it is very likely that their function is to interact with carbohydrates of the membrane (9, 13, 27).

So far, binding of radioactively labeled Vip3A proteins to BBMV has only been performed with ¹²⁵I-Vip3Aa and using partially purified Vip3 proteins as competitors (12, 33–37). In the present work, we have optimized the binding assay, which has allowed us to show ¹²⁵I-Vip3Af specific binding to Spodoptera exigua and Spodoptera frugiperda BBMV and use highly purified Vip3 and Cry proteins to test for shared binding sites. In addition, the use of Ala mutants of Vip3Af that yield truncated versions of the protein has allowed the identification of the Vip3Af domains involved in the specific binding. The role of the domains involved in the specific binding has also been confirmed ex vivo with Sf21 cells.

RESULTS

Specific binding of ¹²⁵I-Vip3Af to S. exigua and S. frugiperda BBMV.

Only marginal specific binding of ¹²⁵I-Vip3Af could be obtained under the conditions described with radiolabeled Vip3Aa or Vip3Ca (20 mM Tris, 150 mM NaCl, 1 mM MnCl₂, 0.1% bovine serum albumin [BSA], pH 7.4) (12, 33). Therefore, new conditions were explored by testing the influence of NaCl and a blocking agent (none, BSA, or membrane blocking agent [MBA]) on S. frugiperda BBMV (Table 1). The results revealed that NaCl had a strong negative effect on the specific binding. Leaving out this salt in the assay allowed us to observe specific binding of ¹²⁵I-Vip3Af, with the best result using BSA as the blocking agent. All subsequent binding assays were performed in the absence of NaCl.

TABLE 1.

Effect of NaCl, MnCl₂, and type of blocker on the total and specific binding of trypsin-treated ¹²⁵I-Vip3Af to S. frugiperda BBMV^a

Binding condition with 20 mM Tris-HCl buffer at pH 7.4			125I-Vip3Af binding (% of input)		% of specific binding over total
NaCl (mM)	Divalent cation (1 mM)	Blocker (0.1%)	Specificb	Total
150			0	9.8	0
150	MnCl2		0	4.8	0
150	MnCl2	BSA	0.1	6.9	1.4
150	MnCl2	MBA	0.4	3.1	12.9
0			15.5	45.2	34.3
0	MnCl2		20.2	50.1	40.3
0	MnCl2	BSA	18.6	28.6	65.1
0	MnCl2	MBA	9.1	23.8	38.4

^aAssays were performed with trypsin-treated 0.37 nM labeled ¹²⁵I-Vip3Af and 0.1 mg/ml BBMV proteins.

^bSpecific binding was determined by adding 370 nM excess of unlabeled Vip3Af (affinity purified by HisTrap FF column).

Figure 1 shows the specific binding of ¹²⁵I-Vip3Af at increasing concentrations of BBMV from S. exigua (Fig. 1A) and S. frugiperda (Fig. 1B). Around 30% of the ¹²⁵I-Vip3Af used in the assay bound to the BBMV, of which approximately 50% was specific binding. SDS-PAGE separation of the proteins in the S. frugiperda BBMV pellet after incubation of ¹²⁵I-Vip3Af and subsequent autoradiography revealed two radioactive peptides bound to the S. frugiperda BBMV: one of 65 kDa (which corresponded to domains II to V of the protein) and another of approximately 19 kDa (domain I) (Fig. 1C). These two bands were remarkably attenuated when the assay was carried out in the presence of an excess of unlabeled Vip3Af. Both the binding curve and the autoradiography showed that a substantial proportion of the nonspecific binding was due to precipitation of Vip3Af or binding to the walls of the assay tubes (radioactivity in the absence of BBMV).

FIG 1.

Specific binding of ¹²⁵I-Vip3Af to S. exigua (A) and S. frugiperda (B). Shown is binding of ¹²⁵I-Vip3Af at increasing concentrations of BBMV; specific binding is the difference between total and nonspecific binding. The nonspecific binding was calculated using an excess (370 nM) of unlabeled Vip3Af toxin (purified by HisTrap FF column). ●, total binding; ○, nonspecific binding. (C) Autoradiography of ¹²⁵I-Vip3Af in different reaction samples. Pellets obtained after centrifugation of the reaction mixture were subjected to SDS-PAGE and exposed to X-ray film. Lanes: 1, labeled toxin used in the assay (input); 2, sample without BBMV (precipitated protein or protein bound to the tube walls); 3, pellet of 0.1 mg/ml S. frugiperda BBMV (total binding); 4, pellet of 0.1 mg/ml S. frugiperda BBMV in the presence of 885 nM unlabeled Vip3Af (nonspecific binding).

Competition binding assays with ¹²⁵I-Vip3Af and Vip3 and Cry1 proteins.

Trypsin-treated Vip3 and Cry1 proteins were used at increasing concentrations to compete with the binding of ¹²⁵I-Vip3Af (Fig. 2). The analysis of the homologous curve (using Vip3Af as competitor) yielded equilibrium dissociation constants (K_d) of 71 ± 7 and 88 ± 2 nM for S. exigua and S. frugiperda, respectively, with a concentration of binding sites (R_t) of 354 ± 32 and 350 ± 21 pmol/mg BBMV protein for S. exigua and S. frugiperda, respectively (Table 2).

FIG 2.

Heterologous competition of Vip3 and Cry1 proteins with ¹²⁵I-Vip3Af. Curves represent binding of ¹²⁵I-Vip3Af at increasing concentrations of unlabeled competitor, using BBMV from S. exigua (A) or S. frugiperda (B). Data points represent the mean and standard error from three biological repeats with duplicated data points.

TABLE 2.

K_d and R_t of Vip3Af and the truncated DI-III molecule with BBMV from S. exigua and S. frugiperda

Protein(s)	Spodopterda exigua a		Spodopterda frugiperda a
	Kd (nM)	Rt (pmol/mg)	Kd (nM)	Rt (pmol/mg)
Vip3Af	71 ± 7	354 ± 32	88 ± 2	350 ± 21
DI-III	72 ± 1	291 ± 6	94 ± 1	426 ± 14

^aResults represent the mean ± standard error (SEM) of the results from four to six replicates.

Binding assays of ¹²⁵I-Vip3Af in the presence of unlabeled heterologous competitors were carried out to evaluate whether Vip3Af shares binding sites with other Vip3 or Cry1 proteins. No competition was observed when Cry1Ac or Cry1Fa was used as a competitor (Fig. 2). In contrast, Vip3Aa and Vip3Ca were able to compete for the ¹²⁵I-Vip3Af binding. In the case of Vip3Aa, the competition curve completely overlapped that of Vip3Af, indicating that both proteins bound to the same sites and with the same affinity to BBMV of the two Spodoptera species. However, Vip3Ca could not displace completely the ¹²⁵I-Vip3Af binding, which means that Vip3Ca only recognizes part of specific sites of Vip3Af.

Competition binding assays with ¹²⁵I-Vip3Af and truncated molecules.

To determine the involvement of the different Vip3Af domains in the specific binding, Vip3Af truncated molecules were tested in competition assays with ¹²⁵I-Vip3Af. The results were similar with BBMV from both S. exigua and S. frugiperda (Fig. 3). The truncated molecules DI-III and DI-IV competed with the trypsin-activated Vip3Af protein (DI-V) for the ¹²⁵I-Vip3Af binding sites (Fig. 3A and B). All three molecules have in common that they conserve domains I to III and that they hold the tetrameric conformation. In contrast, the truncated molecules DIV-V or DIV did not displace ¹²⁵I-Vip3Af within biologically relevant concentrations (Fig. 3C and D). Therefore, the results suggest that the minimum portion of Vip3Af required for specific binding is the one containing domains I to III.

FIG 3.

Heterologous competition of truncated Vip3Af molecules with ¹²⁵I-Vip3Af. Curves represent binding of ¹²⁵I-Vip3Af at increasing concentrations of unlabeled competitor, using BBMV from S. exigua (A and C) or S. frugiperda (B and D). As a control for the competitors, unlabeled trypsin-treated Vip3Af was prepared using the same protocol as the truncated molecules and is represented as DI-V in the figures. Data points represent the mean and standard error from two biological repeats with duplicated data points.

Specific binding of the ¹²⁵I-DI-III molecule to S. frugiperda and S. exigua BBMV.

To further support the role of Vip3Af domains I to III in specific binding, the DI-III molecule was radiolabeled and used in binding assays. Specific binding was confirmed by incubating ¹²⁵I-DI-III (0.4 nM) with increasing concentrations of S. exigua and S. frugiperda BBMV in the presence or absence of an excess of unlabeled DI-III (400 nM) (Fig. 4). At 0.1 mg/ml BBMV, the total binding of ¹²⁵I-DI-III was 30%, of which approximately 60% was specific. Autoradiography after SDS-PAGE separation of the proteins in the S. exigua BBMV pellet showed that both the 37-kDa fragment (which corresponded to domains II to III of the protein) and the 19-kDa fragment bound to the BBMV (Fig. 4B). These two bands were much less intense when the assay contained an excess of unlabeled DI-III.

FIG 4.

Specific binding of ¹²⁵I-DI-III at increasing concentrations of BBMV from S. exigua (A) and S. frugiperda (C). ●, total binding; ○, nonspecific binding. (B) Autoradiography of ¹²⁵I-DI-III in different reaction samples. Lanes: 1, labeled ¹²⁵I-DI-III (input); 2, sample without BBMV (precipitated protein or protein bound to the tube walls); 4, pellet of 0.1 mg/ml S. exigua BBMV (total binding); 3 and 5, pellet of 0.1 mg/ml S. exigua BBMV in the presence of 400 or 1,200 nM unlabeled DI-III, respectively (nonspecific binding).

Competition binding experiments were conducted with ¹²⁵I-DI-III in the presence of increasing concentrations of unlabeled competitors using the truncated molecules (DI-III, DI-IV, DIV-V, and DV) and the trypsin-activated Vip3 proteins (Vip3Af, Vip3Aa, Vip3Ca) (Fig. 5 and 6). The analysis of the homologous curve (using DI-III as competitor) yielded K_d of 72 ± 1 and 94 ± 1 nM for S. exigua and S. frugiperda, respectively, with R_t of 291 ± 6 and 426 ± 14 pmol/mg BBMV protein for S. exigua and S. frugiperda, respectively. These values are very similar to the ones obtained for the Vip3Af protein (Table 2). The results confirmed that domains I to III are critical for the specific binding and that domains IV and V did not act as competitors (Fig. 5A and B). Competition of ¹²⁵I-DI-III with Vip3 trypsin-activated proteins rendered similar results as when using ¹²⁵I-Vip3Af, with Vip3Af and Vip3Aa completely displacing the binding of with ¹²⁵I-DI-III and with Vip3Ca competing only partially (Fig. 6A and B). These results confirm that the specific binding observed in vitro is mainly due to the three most-N-terminal domains and that more than one population of binding sites exists for Vip3Af since Vip3Ca can displace just a fraction of the binding.

FIG 5.

Heterologous competition of truncated proteins with ¹²⁵I-DI-III to BBMV from S. exigua (A) and S. frugiperda (B). Data points represent the mean and standard error from two biological repeats with duplicated data points. Note that DI-V refers to the trypsin-activated Vip3Af purified in the same way as the truncated fragments.

FIG 6.

Heterologous competition of Vip3 proteins with ¹²⁵I-DI-III to S. exigua (A) and S. frugiperda (B). Data points represent the mean and standard error from two biological repeats with duplicated data points.

Toxicity of Vip3Af and its truncated molecules to Sf21 cells.

To determine the appropriate conditions for the assays, the isoelectric point precipitation (Ipp)-purified Vip3Af protein was first tested both as a protoxin and as the trypsin-activated form, at different concentrations. The results showed that the protoxin had a minor effect on the viability of Sf21 cells, which was only detectable after 48 h of exposure at the highest concentration tested (Fig. 7A and B), whereas the trypsin-activated Vip3Af was toxic in a dose-dependent manner. Moreover, the loss of cell viability was slightly higher when Sf21 cells were exposed to the activated protein for 48 h than when cells were exposed for 24 h.

FIG 7.

Morphology (A) and cell viability (B and C) of Sf21 cells exposed to Vip3Af and its truncated molecules for 48 h. Cell viability assays were repeated more than three times, each time including at least three replicates. The bars represent the mean and standard error of the mean. (D) Specific binding of ¹²⁵I-DI-III to Sf21 cells. ●, total binding; ○, nonspecific binding.

Based on the above results, the truncated molecules were tested at 100 μg/ml both as Ipp purified and as chromatography purified. As a control, the trypsin-activated Vip3Af protein (DI-V) was purified in the same way as the truncated molecules and tested in parallel. The results showed that the molecule DIV-V had basically no effect on the viability of Sf21 cells (Fig. 7C), even when the concentration was increased to 300 μg/ml (data not shown), whereas the DI-III and DI-IV molecules were as active as the trypsin-activated Vip3Af protein (DI-V). These results indicate that the fragment containing DI-III is the minimum Vip3Af fragment required for toxicity to Sf21 cells.

Specific binding assays of ¹²⁵I-DI-III to Sf21 cells.

Since the DI-III truncated molecule was toxic to Sf21 cells, we set out to demonstrate specific binding of ¹²⁵I-DI-III to these cells using the same approach as with BBMV. Figure 7D shows that ¹²⁵I-DI-III binds specifically and in a concentration-dependent manner to Sf21 cells. Therefore, the DI-III fragment is also the minimum Vip3Af fragment required for binding to Sf21 cells.

DISCUSSION

The first steps toward understanding the mode of action of Vip3 proteins were given at the time of their discovery from Bt in 1996 (38). To date, although more than 101 Vip3 proteins have been reported (39), their mode of action remains somewhat elusive (9, 40, 41), though it is well accepted that their activity requires binding to membrane receptors in the midgut of target insects (12, 16, 42, 43). Several insect proteins have been proposed as functional receptors for Vip3Aa (17–20). However, very little is known regarding how Vip3 proteins interact with the membrane receptors and which Vip3 domains are involved.

So far, evidence that Vip3 proteins bind specifically to the brush border membrane of the midgut epithelial cells has been shown, almost exclusively, in studies working with the Vip3Aa protein. Specific binding of Vip3Aa has been shown in Spodoptera littoralis, S. frugiperda, A. ipsilon, Heliothis virescens, Helicoverpa zea, and Helicoverpa armigera (12, 16, 34, 35, 42–44). Only a few studies have shown binding of other Vip3 proteins, Vip3Af and Vip3Ca among them, and except for the latter (33), always indirectly as heterologous competitors of Vip3Aa or in qualitative or semiquantitative assays (12, 15, 36). However, direct binding of radioactively labeled Vip3Af had never been reported before. By excluding NaCl from the binding assay buffer, we have been able to show specific binding of Vip3Af to BBMV from S. exigua and S. frugiperda larvae and perform competition binding assays to determine whether Vip3Af binding sites are shared with other Bt toxins. Autoradiography of the BBMV pellet after the binding assay showed that both the 19- and 65-kDa fragments bind specifically (because their binding is competed by an excess of unlabeled protein) to the BBMV (Fig. 1), in agreement with previous reports that showed that these two fragments remain strongly associated after trypsin activation (12, 13).

Competition binding studies are a potent tool that provides information on the potential for cross-resistance among Bt toxins since alteration of binding sites is the main mechanism of resistance to Cry proteins (45–47). All studies performed with insects resistant to either Cry or Vip3A proteins have shown lack of significant cross-resistance between these two families of proteins (37, 48–54), which is supported by the fact that they do not share binding sites (9, 12, 33, 36, 55). Our results with S. exigua and S. frugiperda support the lack of shared binding sites between Cry1 and Vip3 proteins and the occurrence of common sites for Vip3 proteins (Fig. 2). Our results also expand the binding site model of Vip3 proteins (15, 33) in that they show that Vip3Ca is not able to compete for all Vip3Af binding sites, suggesting that Vip3Af has some binding sites to which Vip3Ca does not bind (Fig. 8).

FIG 8.

Proposed binding site model for Vip3 and Cry1 toxins in Spodoptera spp.

Four independent groups have recently reported the 3D structure of Vip3Aa and Vip3B proteins (13, 25, 28, 56), which has firmly established 5 structural domains that had been previously proposed (29). The structure of the trypsin-activated Vip3Aa has shed light on the long-time puzzling observation that the 19- and 65-kDa fragments remain strongly associated after trypsin action (13). The 3D structures confirmed that Vip3 protoxins spontaneously form homotetramers in solution, which adopted a “pyramid-shaped” structure. Upon protease activation, and probably triggered by binding to a receptor, the tetrameric molecule suffers a drastic change, conferring on it a new conformation with a “syringe-like” structure, with the “needle” being a four-helix coiled coil involving domain I from the four monomers (13).

By the use of Vip3Af Ala mutants unstable with trypsin, we have been able to generate and purify Vip3Af truncated molecules with a different domain composition. Instead of using cloned fragments, we preferred to use the alternative approach with the aim of maintaining the original 3D structure of the remaining part of the truncated molecule, especially the tetrameric structure. From three Vip3Af Ala mutants (W552A, I699A, and F229A), we produced the DI-III and DI-IV molecules, which maintain the tetrameric structure (29), and the DIV and DIV/V molecules, which contain the carbohydrate-binding modules (13, 25, 28, 56) and cannot form tetrameric structures in the absence of the rest of the protein (29). In competition binding assays with BBMV, we could demonstrate that the truncated molecules containing domains I to III were able to compete for binding with ¹²⁵I-Vip3Af, whereas domain IV or IV/V was not (Fig. 3). These results were confirmed by labeling the DI-III truncated molecule, which showed specific binding to BBMV (Fig. 4). Only those molecules retaining domains I to III were able to compete with the ¹²⁵I-DI-III molecule for BBMV binding (Fig. 5). As with ¹²⁵I-Vip3Af, Vip3Af and Vip3Aa completely displaced binding of the DI-III truncated molecule, whereas Vip3Ca competed only partially (Fig. 6).

Specific binding of the ¹²⁵I-DI-III molecule to Sf21 cells was also observed by performing binding assays in the presence and absence of unlabeled competitor. Our results are in agreement with those reported with cloned Vip3Aa fragments using fluorescence-based cell binding assays with Sf9 cells (28). The cloned DI-III and DII-III fragments displayed total binding to the cells similar to that of the full-length Vip3Aa (protoxin), whereas the binding of the cloned DIII or DI-II fragments was less efficient, and no binding was observed for the DIV-V fragment. We have also tested the functional role of the truncated molecules by measuring their toxicity to Sf21 cells. In agreement with the binding data, only the trypsin-activated Vip3Af protein (DI-V) and the truncated molecules maintaining domains I to III were toxic (Fig. 7C), indicating that the binding observed to the cells was functional. The results from the in vitro binding to BBMV and the ex vivo binding to Sf21 cells, along with the functional analysis (toxicity to Sf21 cells) and results from other authors (28), strongly indicate that domains I to III are critical in maintaining the functional core of the Vip3 proteins. Domain III contains three antiparallel β-sheets that form a β-prism fold strikingly similar to that found in the Cry insecticidal δ-endotoxins (13, 57). Therefore, it is a strong candidate to be the domain interacting with the membrane receptor, either alone or in combination with domain II or domain I plus domain II. Although the C-terminal variable domains IV and V do not seem to play a role in either the in vitro and ex vivo binding or the toxicity to Sf21 culture cells, it has been widely demonstrated that mutations in those domains drastically decrease the Vip3A insecticidal activity and that their presence is required for in vivo toxicity (26, 29–32, 58, 59). Our results point out that these domains are critical for in vivo activity (insecticidal activity) but not for ex vivo activity (cell toxicity to Sf21 cells). The difference in the need for domains IV and V in these two scenarios is probably due to the protective effect of domains IV and V in vivo against lumen proteases, something that poses no problem with the ex vivo experiments. Alternatively, domains IV and V may have a role in bringing the toxin close to the membrane by binding to abundant glycosylated molecules (which may not be present in the Sf21 cells), similarly to the role of aminopeptidase N with Cry proteins (60). Therefore, further research is needed to determine the role of these domains in vivo.

In conclusion, the results in this study provide evidence of the critical role of the N-terminal domains (domains I to III) in the mode of action of Vip3Af and probably all Vip3 proteins. The role of the highly variable C-terminal domains (domains IV and V) remains elusive, though they are required for full toxicity in vivo. It is likely that their role in vivo is to increase the stability of the tetrameric structure and to bind to glycosylated molecules in the apical membrane to help the Vip3 bind to the specific functional receptors.

MATERIALS AND METHODS

Source of proteins and expression.

The insecticidal proteins used in the present work were obtained from different sources. The Vip3Af1 protein (NCBI accession no. CAI43275 [here Vip3Af]), three Vip3Af Ala mutant proteins (Vip3Af-F229A, -W552A, and -I699A), and the Vip3Ca2 protein (NCBI accession no. AEE98106 [here Vip3Ca]) were overexpressed in Escherichia coli WK6 carrying the expression vector pMaab 10 (kindly supplied by Bayer CropScience N.V., Ghent, Belgium [currently BASF Belgium Coordination Center–Innovation Center Gent]). Vip3Aa16 (here Vip3Aa) was prepared from recombinant E. coli BL21 expressing the vip3Aa16 gene (accession no. AY739665) (61). Protein expression and lysis of Vip3Af, its mutants, and Vip3Ca were carried out under the conditions described previously (15), whereas expression of Vip3Aa was performed as described elsewhere (61). Cry1Ac and Cry1Fa were obtained from recombinant B. thuringiensis strains EG11070 and EG11069, respectively (from Ecogen, Inc., Langhorn, PA). Crystal purification and solubilization were performed as described before (62).

Purification of Vip3 and Cry1 proteins.

To be used for radiolabeling, the Vip3Af protein was subjected to isoelectric point precipitation (Ipp) and to anion-exchange chromatography. After cell lysis, the pH of the Vip3Af lysate was lowered with 0.1 M acetic acid to pH 5.6. After 10 min of centrifugation at 16,100 × g at 4°C, a portion of the pellet was dissolved with buffer Tris-NaCl (20 mM Tris, 150 mM NaCl, pH 9.0) and dialyzed against 20 mM Tris at pH 9.0 overnight. Afterwards, the protein was filtered prior to anion-exchange purification in a HiTrap Q HP column using an ÄKTA explorer 100 chromatography system (GE Healthcare, United Kingdom). Proteins were eluted with a 100-ml linear gradient (0 to 80%) of 1 M NaCl. The collected 1-ml fractions were subjected to SDS-PAGE. Fractions containing Vip3Af were stored, and the protein concentration was quantified by Bradford assay. The fraction with the maximum concentration of Vip3Af was used for radiolabeling.

The Vip3 proteins to be used as competitors in binding assays were purified from crude extracts (lysed cells) by metal-chelate affinity chromatography using 1 ml HisTrap FF columns (GE Healthcare) (63). The fractions containing the Vip3 protein were pooled and dialyzed overnight against a mixture of 20 mM Tris, 150 mM NaCl, and 5 mM EDTA at pH 8.6 and stored at −80°C. The purified Vip3 proteins were trypsin treated (5% [wt/wt] trypsin from bovine pancreas [Sigma T8003; Sigma-Aldrich, St. Louis, MO, USA]) at 37°C for 1 h and centrifuged at 16,100 × g for 10 min at 4°C, and the supernatant was quantified by Bradford assay (64) (see Fig. S1 in the supplemental material).

Trypsin-activated Cry1Ac and Cry1F proteins were further dialyzed in 20 mM Tris-HCl (pH 9) and filtered prior to being purified by anion-exchange chromatography in a HiTrap Q HP column as described elsewhere (62) (Fig. S1).

For cell toxicity assays, Vip3Af was subjected to Ipp as described above. Then, a portion of the pellet was dissolved with a mixture of 20 mM Tris and 150 mM NaCl at pH 9.0 and dialyzed against the same buffer overnight. The concentration of the Vip3Af protoxin was determined, before trypsin activation, by densitometry after SDS-PAGE using bovine serum albumin (BSA) as standard and the TotalLab 1D v13.01 software. The Ipp-purified protein was flash frozen in liquid nitrogen and stored at −80°C until used.

Generation and purification of truncated Vip3Af molecules.

According to our previous results, the Vip3Af Ala F229A, W552A, and I699A mutations destabilize the protein, making some sites more accessible to trypsin and generating fragments containing different combinations of the structural domains (26, 29). We have used the trypsin treatment approach to generate Vip3Af truncated molecules with a different domain composition, which in principle, would maintain the original 3D structure like the full-length protein since they would be derived from the original folded protoxin.

The main fragments resulting from treatment of F229A with trypsin were the 17- and 27-kDa fragments (29). The 17-kDa fragment corresponds to domain IV and the 27-kDa fragment to domains IV and V (26, 29). To obtain pure preparations of these fragments, the F229A protoxin was first purified by the HisTrap FF column and dialyzed in Tris buffer (20 mM Tris-HCl, pH 8.6) overnight. The protoxin was then incubated with trypsin (5% [wt/wt], 37°C, 1 h), and the fragments were separated by anion-exchange chromatography in a HiTrap Q HP column (5-ml bed volume), equilibrated in the same dialysis buffer, in an ÄKTA explorer 100 chromatography system, and eluted with a 100-ml linear gradient (0 to 80%) of 1 M NaCl. Individual peaks containing the 17-kDa fragment (DIV) and the 27-kDa fragment (DIV-V) were collected, subjected to SDS-PAGE, quantified by Bradford assay (64), frozen in liquid nitrogen, and stored at −80°C (see Fig. S2 in the supplemental material). Fractions A5 and A9 were used for competition analyses and for cell toxicity assays.

The incubation of W552A and I699A mutants with trypsin generates, in addition to the N-terminal 19-kDa fragment (from amino acids 12 to 198, corresponding to domain I) (26, 29), a main fragment of around 38 kDa (for the W552A mutant) or one of around 53 kDa (for the I699A mutant), which along with the 19-kDa fragment maintain the tetrameric structure of the protein (29). The 38-kDa fragment consisted mostly of domains II and III (29). The 53-kDa fragment was previously identified as containing domains II, III, and IV (26).

To obtain pure preparations of the 19- plus 38-kDa (DI-III) and 19- plus 53-kDa (DI-IV) truncated molecules, the W552A and I699A mutants were first purified by Ipp (pH 5.6), subjected to trypsin treatment (5% [wt/wt], 37°C, 1 h), and finally purified by gel filtration chromatography in a Superdex 200 column (Superdex 200 10/300 GL; GE Healthcare Life Sciences, Uppsala, Sweden) equilibrated and eluted with Tris-NaCl buffer (20 mM Tris-HCl, 150 mM NaCl, pH 8.6). Under these conditions, the tetrameric molecules can be separated from small fragments that do not form oligomers (see Fig. S3 in the supplemental material). Individual peaks containing the pure fragments were collected, subjected to SDS-PAGE, quantified by Bradford assay (64), frozen in liquid nitrogen, and stored at −80°C. Fractions A11 and A12 were used for competition analyses and for cell toxicity assays. The Vip3Af protein was also subjected to the same protocol to serve as a control in the assays with the truncated molecules (referred to in these experiments as DI-V). Figure S3 shows the chromatograms and SDS-PAGE of the collected fractions. Fractions A11 and A12 were used for competition analyses and for cell toxicity assays.

BBMV preparation.

Midguts from fifth instar larvae of S. exigua and S. frugiperda were used to prepare BBMV by the differential magnesium precipitation method (65). BBMV were immediately frozen in liquid nitrogen and stored at −80°C until used (less than 1 month). The protein concentration of BBMV preparations was determined by Bradford assay (64) using BSA as a standard.

Radiolabeling of Vip3Af and the DI-III molecule.

Chromatographically purified Vip3Af protoxin (25 μg) or the DI-III molecule (25 μg) was labeled with 0.3 mCi of [¹²⁵I]NaI by the chloramine T method (12, 66). The mixture was passed through a PD10 desalting column (GE Healthcare Life Sciences, United Kingdom) to separate the labeled protein from the excess of free radioactive iodine. The protein peak fractions were collected and stored at 4°C. For optimal results, binding assays were performed within 10 days of protein labeling. The specific activities of labeled proteins were 3.81 mCi/mg for Vip3Af protoxin and 0.75 mCi/mg for the DI-III molecule, respectively.

Binding assays with ¹²⁵I-labeled Vip3Af and the ¹²⁵I-DI-III molecule to BBMV.

Prior to being used in binding assays, radiolabeled Vip3Af protoxin was trypsin treated (5% [wt/wt], 37°C, 1 h) and stored at 4°C until used. BBMV, which had been stored at −80°C, were thawed on ice and centrifuged for 10 min at 16,000 × g at 4°C, and then the supernatant was removed and the pellet resuspended in binding buffer (20 mM Tris, 1 mM MnCl₂, 0.1% BSA, pH 7.4). The binding assay consisted of incubating the ¹²⁵I-Vip3Af (0.37 nM) or ¹²⁵I-DI-III (0.40 nM) with 0.1 mg/ml BBMV (except when different concentrations of BBMV were tested) for 1 h at room temperature (RT) in a 0.1-ml final volume of binding buffer. The reaction was stopped by centrifuging the tubes at 16,100 × g for 10 min at 4°C, the supernatant was removed, and the pellet was washed twice (centrifuged at 16,100 × g for 5 min at 4°C) with 500 μl of cold binding buffer. The radioactivity in the final pellet was measured in a model 2480 Wizard2 gamma counter. An excess of unlabeled protein (370 nM trypsin-treated Vip3Af or 400 nM DI-III molecule) was used to estimate the nonspecific binding.

To search for the optimal conditions for Vip3Af specific binding, a fixed amount of trypsin-treated ¹²⁵I-Vip3Af (0.37 nM) was used with 0.1 mg/ml BBMV to test the effect of NaCl concentration, MnCl₂, and the type of blocking agent: BSA or membrane blocking agent (MBA [GE Healthcare, United Kingdom]).

Autoradiography of ¹²⁵I-Vip3Af or ¹²⁵I-DI-III bound to BBMV was performed by resuspending the binding assay pellets in 10 μl Milli-Q water and mixing them (2:1 [vol/vol]) with loading buffer (0.2 M Tris-HCl, pH 6.8, 1 M sucrose, 5 mM EDTA, 0.1% bromophenol blue, 2.5% SDS, and 5% β-mercaptoethanol). After heating at 99°C for 5 min, the mixture was subjected to 12% SDS-PAGE. When finished, the gel was dried (at 50°C for 1 h) and exposed to X-ray film.

Cell culture maintenance.

Spodoptera frugiperda-derived Sf21 cells were cultured as a monolayer at 25°C in Gibco Grace’s medium (1×) (Life Technologies) supplemented with 10% heat-inactivated fetal bovine serum (FBS). Routine culture was performed in T25 flasks (Nunc), and cells were passaged into fresh medium once a week.

Toxicity assays with Vip3Af and truncated Vip3Af molecules.

For cell toxicity assays, Sf21 cells were suspended in Grace’s medium (without FBS) and plated in flat-bottom 96-well enzyme-linked immunosorbent assay (ELISA) plates at ca. 70% confluence. A total volume of 100 μl of cell suspension (6 × 10⁵ cells/ml) was added per well, and the plates were left standing at 25°C for at least 45 min. Cell viability assays were performed using the MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assay (67) after exposure of cells to Vip3Af proteins. At least five concentrations of the Ipp-purified Vip3Af protein (10 μl), either as a protoxin or as a trypsin activated, were tested. The 20 mM Tris–150 mM NaCl buffer at pH 9.0 was used as a negative control and 2% Triton X-100 as a positive control. After 24 or 48 h of incubation at 25°C, cell viability was measured using the CellTiter 96 AQueous One Solution reagent (Promega, Madison WI) following the manufacturer’s protocol. Briefly, 20 μl of the reagent was added to each well, and then the plate was further incubated for 2 h at 25°C. After incubation, absorbance was measured at 490 nm (Infinite m200; Tecan, Maennedorf, Switzerland). The percentage of viability was calculated in relation to the absorbance of the cells treated with buffer (defined as 100% viable) and that of the cells treated with Triton X-100 (defined as 0% viable) (68). Duplicates of each condition were performed in each assay, and assays were replicated at least three times.

Vip3Af truncated molecules (DI-III, DI-IV, DIV-V, DV), purified by Ipp or by chromatography, were tested for cell toxicity as described above, though only at a single concentration (100 μg/ml). As a control, Vip3Af was purified after trypsin activation following the same methodology used for the truncated molecules and tested in parallel (referred to in these experiments as DI-V). Duplicates of each condition were performed in each assay, and assays were replicated at least twice.

Binding of ¹²⁵I-DI-III to Sf21 cells.

Specific binding of the labeled DI-III was analyzed on Sf21 cells. Prior to binding assays, Sf21 cells were detached and recovered by centrifugation at 500 × g for 5 min at RT and then washed twice with binding buffer. The pellet of cells was gently suspended in binding buffer to a concentration of approximately 4 × 10⁷ cells/ml, calculated by the Countess automated cell counter (Invitrogen). Total binding was determined by incubating different amounts of cells with ¹²⁵I-DI-III (0.4 nM) in a final volume of 0.1 ml in binding buffer. The same experiment was done with an excess of unlabeled DI-III (400 nM) added to each tube to calculate the nonspecific binding. After 1 h at RT, the reaction was stopped by centrifuging at 500 × g for 10 min at RT, and the pellets were washed with 500 μl of binding buffer twice. Radioactivity was measured in a model 2480 Wizard2 gamma counter. Two replicates were performed.