Improvement of Crystal Solubility and Increasing Toxicity against Caenorhabditis elegans by Asparagine Substitution in Block 3 of Bacillus thuringiensis Crystal Protein Cry5Ba

Abstract

The crystal proteins from Bacillus thuringiensis are widely used for their specific toxicity against insects and nematodes. The highly conserved sequence blocks play an important role in Cry protein stability and flexibility, the basis of toxicity. The block 3 in Cry5Ba subfamily has a shorter sequence (only 12 residues) and more asparagine residues than that of others which harbor about 48 residues but only one asparagine. Based on the theoretical structure model of Cry5Ba, all three asparagines in block 3 are closely located in the interface of putative three domains, implying their probable importance in structure and function. In this study, all three asparagines in Cry5Ba2 block 3 were individually substituted with alanine by site-directed mutagenesis. The wild-type and mutant proteins were overexpressed and crystallized in acrystalliferous B. thuringiensis strain BMB171. However, the crystals formed in one of the mutants, designated N586A, abnormally disappeared and dissolved into the culture supernatant once the sporulation cells lysed, whereas the Cry5Ba crystal and the other mutant crystals were stable. The mutant N586A crystal, isolated from sporulation cells by the ultrasonic process, was found to be easily dissolved at wide range of pH value (5.0 to 10.0). Moreover, the toxicity assays showed that the mutant N586A exhibited nearly 9-fold-higher activity against nematodes and damaged the host's intestine more efficiently than the native Cry5Ba2. These data support the presumption that the amide residue Asn586 at the interface of domains might adversely affect the protein flexibility, solubility and resultant toxicity of Cry5Ba.

INTRODUCTION

The spore-forming bacterium Bacillus thuringiensis can produce a large family of insecticidal and nematicidal proteins (43, 47). These proteins, named Cry proteins, form parasporal crystals during the sporulation phase in cell growth cycle (22). Most Cry proteins exhibit significant similarity in three-dimensional (3D) structure and mode of action, although they have considerable differences in protein sequence and target specificity (40). Proved as pore-forming toxins, the three-domain Cry proteins lyse midgut epithelial cells by inserting into the cell membrane of the insect midgut, resulting in pore formation and host death (6). The transformation of Cry proteins from inactive crystal protoxin to cytotoxic toxin is a multistep process. When dissolved and activated upon ingestion by susceptible animal, the Cry proteins bind to specific receptors on the midgut epithelium, leading to toxin oligomerization, membrane insertion, and resulting pore structure formation (7).

Primary- and tertiary-structure analysis of Cry proteins has provided considerable insight into their structural and functional action. To date, eight 3D structures have been elucidated: Cry3Aa (26), Cry1Aa (18), Cry1Ac (12), Cry3Bb (15), Cry2Aa (33), Cry4Ba (4), Cry4Aa (5), and Cry8Ea (21). These toxins share highly similar three-domain structures, in which domain I, an α-helical bundle, is thought to be involved in membrane insertion and pore formation, and domains II and III, a β-prism and β-sandwich, respectively, are believed to be the major determinants of receptor binding (40). Complete sequence alignment of Cry proteins revealed five highly conserved blocks in the N-terminal half (43). Based on spatial structure of Cry proteins, the five blocks were found to lie in the center of individual domain or interface of three domains (18), implying their putative involvement in interdomain contacts and balance of protein stability and flexibility.

The role of these conserved blocks was emphasized by Cry protein mutagenesis studies. It has been demonstrated that the Cry3A block 1 was involved in both host pore formation and protein protection from enzymatic proteolysis (16). The substitution of amino acid residues in Cry4A block 5 resulted in decreased production but did not affect the insecticidal activity (35). In scarabaeid beetle-specific Cry8Ca, whole sequence of any of the five blocks is necessary for activity against cupreous chafer (32).

The nematicidal Cry5Ba subfamilies are believed to be a potential resource for genetically modified plant breeding to control parasitic nematodes (24, 27). A novel block 3 conserved in Cry5Ba subfamilies was identified as a short sequence that consists of only 12 residues, whereas the corresponding block in most of other crystal proteins contain about 48 residues (43). In the theoretical 3D structures of Cry5Ba (49) and Cry5Aa (50), the third blocks were all situated in the core of the molecule and interface of putative three domains (see Fig. S1 in the supplemental material), indicating their probable involvement in protein structural and functional role.

Another characteristic of block 3 in Cry5Ba subfamily is the richness of the conserved asparagine residue, which harbors two to three asparagines in a total of 12 residues, whereas only one asparagine present in the third block of other crystal proteins (see Fig. 1). As a kind of amide amino acid, the asparagine residue exhibits polar behavior and tends to an form intra- or intermolecular hydrogen bond (9) and plays an important role in structural and functional action (41). It was reported that asparagine 175 of connexin32 is a critical residue for docking and forming functional heterotypic gap junction channels with connexin26 (34). As for Cry proteins from B. thuringiensis, specific substitutions of Asn166 within the α4-α5 loop in domain I of Cry4B almost completely abolished the toxicity against mosquito larvae, suggesting that the polarity at Asn166 plays a crucial role in the larvicidal activity (23). The residues Asn576 and Asn584 in β20-β21 loop of the domain III of Cry1Ac were demonstrated to be on the inside of the molecular and considered critical to the stability of the protein structure (29).

Fig 1.

Multiple sequence alignment of block 3 of Cry5Ba subfamily and their homologous proteins. (A) Highly conserved block 3 in Cry8Ea, Cry3Bb, Cry3Aa, and Cry1Aa. (B) Novel block 3 highly conserved in Cry5Ba subfamilies. Arrows indicate the three nearly successive asparagine residues N586, N588, and N599 in block 3 of Cry5Ba.

In the present study, to explore the role of asparagines in the special block 3 of Cry5Ba toxin, we introduced single asparagine substitution to block 3 of Cry5Ba2. The Cry5Ba2 protein and its mutants N586A, N588A, and N589A were all overexpressed and accumulated to form crystals in B. thuringiensis. However, the mutant N586A improved the crystal inclusion solubility across a wide range of pHs and enhanced the protein toxicity against the nematode Caenorhabditis elegans. Our findings provide a novel strategy to modify the Cry proteins for efficient control of this pest.

MATERIALS AND METHODS

Cry5Ba conserved blocks analysis.

By protein BLAST with Cry5Ba2 protein sequence in NCBI PDB database, the four proteins Cry8Ea (PDB code 2QKG), Cry3Bb (PDB code 1JI6), Cry3Aa (PDB code 1DLC), and Cry1Aa (PDB code 1CIY) were found to be most homologous to Cry5Ba. By multiple sequence alignment using CLUSTAL W (25), the five conserved blocks were identified in the subfamily of Cry5Ba, together with the four homologous proteins, as reported by Schnepf et al. (43). The differences of block 3 between them was found and further analyzed. Given that the theoretical 3D structure model of Cry5Ba1 had been predicted by Xia et al. (49), the PDB file (PM0075036) was downloaded from CASPUR PMDB database and used for spatial localization of the block 3 and its asparagine residues in the molecule by using Swiss PdbViewer 4.01 (19).

Site-directed mutagenesis.

The mutation reaction was performed by using a stepwise method, PCR-driven overlap extension (8) with the primers described in Table 1. In the first step, with N586A as example, primer pairs F1/R1(586) and F2(586)/R2 were, respectively, used to amplify the overlap fragments covering the mutation with native plasmid pBMB0215 (20) as a template, which harbor the native cry5Ba gene in vector pHT304. The two amplified fragments were purified and then mixed as overlapped templates in the second step, where primer pair F1/R2 was also added to increase the efficiency of amplification. The fully amplified mutated fragments were respectively purified, digested with PstI and BglII, and cloned back into modified plasmid pBMB0215, in which corresponding native fragment was cut off by PstI and BglII in advance. All mutation plasmids, verified by restriction enzyme mapping and DNA sequencing, were transferred into acrystalliferous B. thuringiensis strain BMB171 by electroporation for protein production.

Table 1.

Primers used for mutagenesis of cry5Ba by PCR-driven overlap extension

Primer	Sequence (5′–3′)	Length (bp)	Underlined sequence
F1	CCTCTAGAGTCGACCTGCAGGCATG	25	PstI site
R2	TAAAAAGAGATCTGAAGAACCTTGG	25	BglII site
R1(586)	TTAACTGCATTGTTACCAGCTATCCATTCGCGAACAAG	38	Mutant site
F2(586)	CTTGTTCGCGAATGGATAGCTGGTAACAATGCAGTTAA	38	Mutant site
R1(588)	AAAGTTTAACTGCATTGGCACCATTTATCCATTCGC	36	Mutant site
F2(588)	GCGAATGGATAAATGGTGCCAATGCAGTTAAACTTT	36	Mutant site
R1(589)	TTAGAAAGTTTAACTGCAGCGTTACCATTTATCCATTC	38	Mutant site
F2(589)	GAATGGATAAATGGTAACGCTGCAGTTAAACTTTCTAA	38	Mutant site

Protein expression and crystallization detection.

The B. thuringiensis cells were allowed to culture in liquid PM medium (46) at 28°C until approximately half (about at 40 h) or nearly all (about at 48 h) of the sporulation cells lysed. The cell cultures were collected for protein characterization on protein gels relative to bovine serum albumin standards (30) and crystal observation with phase-contrast microscope (Olympus BX51) as described previously (20). For electron microscopic observation, the crystal-matured prelysed cell cultures were collected and washed by 0.5 M NaCl–10 mM EDTA solution (pH 5.0). The cells were allowed to lyse by suspending them with 0.1 M phosphate buffer solution (pH 5.0, 6.0, and 7.0) at 4°C for 24 h. For scanning electron microscopy (SEM) with a Quanta200 (FEI), the lysed cell samples were treated by following the methods described by Shao et al. (44); for transmission electron microscopy (TEM) using a Tecnai G2 20 (FEI), the lysed cell samples were treated as described by Bailey-Smith et al. (2).

The temporal protein expression of N586A and Cry5Ba in BMB171 was further analyzed. The B. thuringiensis cell cultures in liquid PM medium were sampled every 4 h from 24 to 48 h. After centrifugation for 5 min at 10,000 × g, the culture precipitates were prepared for inclusion proteins detection in SDS-PAGE gels, and the culture supernatants were treated with trichloroacetic acid-acetone for soluble protein precipitation and concentration according to the method of Wessel et al. (48).

Crystal solubility assessment.

The cell cultures in liquid PM medium (46) were collected when the crystals were nearly mature in prelysed sporulation cells about at 36 h. After centrifugation for 5 min at 10,000 × g, the intact cell precipitates were suspended with 0.5 M NaCl–10 mM EDTA solution at pH 5.0 and artificially crushed by ultrasonic processor (Vibra-Cell VCX130). The dissolution reaction was carried out by incubating the released crystals in 50 mM Na₂CO₃ buffer with different pHs (5.0 to 10.0) at 4°C for 5 h. After centrifugation for 15 min at 10,000 × g, the soluble proteins in supernatant were prepared for characterization using SDS-PAGE. In addition, for the next spectroscopy and bioassay, the dissolved proteins were further purified by isoelectric precipitation in Na₂CO₃-sodium acetate buffer at pH 5.0 (20) and quantitatively determined based on the methods of Lowry et al. (28).

CD spectroscopy.

Circular dichroism (CD) spectra of N586A and Cry5Ba were collected with a J-810 CD spectrometer (Jasco Corp., Japan). The purified proteins were quantified by UV and visible spectrophotometer UV-9200 and diluted to a concentration of 0.1 mg/ml in 2 mM HEPES buffer. The data shown were based on scanning from 195 to 260 nm at 1.0-nm steps and averaged from three scans.

Nematode toxicity assay.

The growth assay was carried out by feeding the first larval (L1)-stage C. elegans with different doses of purified crystal proteins using a 96-well plate according to the method of Bischof et al. (3). Five concentration gradients in each protein and five independent trials in each concentration were used. Each well contained 10 μl of freshly hatched L1 stage animals in M9 buffer (about 15 to 20 worms), 10 μl of 100-fold protein (500, 50, 5, 0.5, or 0.05 μg/ml in 20 mM Heps buffer) or control Heps buffer, 10 μl of Escherichia coli OP50 in S medium (final optical density at 600 nm of ∼0.6), and 70 μl of S medium with a final volume of 100 μl in each well. The pH for all buffers used here was ∼7.0. The peripheral wells in the plate were not used for the toxicity trial but contained 100 μl of sterilized water to preserve moisture. The L1 worms incubated with purified crystal proteins or control buffer were allowed to grow at 20°C for 72 h. The worms were photographed at 100-fold magnification on a phase-contrast microscope (Olympus BX51). The worm images were outlined and measured using the National Institutes of Health Image program ImageJ 1.33, and the worm size was normalized to the average worm size of the control. The Statistical Analysis System (SAS 8.0) was used to generate the statistics on the data.

As for brood size assay, the fourth larval (L4) stage animals, fed with purified crystal protein, were incubated in S medium (3) at 25°C for 72 h. The number of progeny was counted on microscope (Olympus BX51). The brood size was normalized to the average brood size of the control. The data were analyzed by using SAS 8.0 to generate the statistics.

In the intestine damage assay, L4 stage animals were incubated in S medium with 500 ng of purified crystal protein/ml at 20°C for 24, 48, and 72 h, according to the methods of the lethal-dose assay (3). The worm images were photographed in BF mode (visible light) and DAPI mode (autofluorescence; excitation wavelength, 330 to 385 nm; emission wavelength, 420 nm) on Zeiss Scope A1 fluorescence microscope.

Protease sensitivity assay.

The purified crystal proteins were quantitatively determined by the method of Lowry et al. (28) and then treated with trypsin (Promega) at a mass ratio of 0.0004:1 (trypsin/Cry protein). The reaction was performed at 28°C for 10 to 50 min in 0.1 M phosphate-buffered solution (PBS) at pH 8.0.

RESULTS

Alanine substitution of asparagines in block 3 of Cry5Ba.

B. thuringiensis Cry5Ba protein subfamilies have a novel block 3 significantly different from highly conserved group of major Cry proteins (43) (Fig. 1). By multiple sequence alignment, the block 3 of Cry5Ba2 was identified, which has only 12 residues and contains the three nearly successive amide amino acid residues N586, N588, and N599 (Fig. 1B). In theoretical 3D structure of Cry5Ba1 (PM0075036) from the PMDB database (49), we marked the three asparagine residues in block 3 with different colors and found that the three residues were located in interface between domains I, II, and III (see Fig. S1 in the supplemental material). To know the role of the three asparagines in protein structure and nematicidal activity, the asparagines were individually substituted by a small amino acid alanine. The mutants were all transferred into acrystalliferous B. thuringiensis strain BMB171 for crystal protein production.

Protein expression and crystal formation of Cry5Ba and its mutants.

Our previous study (20) showed that overexpressed Cry5Ba protein accumulated in B. thuringiensis sporulation cells to form crystals that would be released from the sporulation mother cell when lysed. To examine crystal formation, Cry5Ba and its mutants were overexpressed in strain BMB171. After culture (40 h), the Cry5Ba protein and its mutants N586A, N588A, and N589A were found to be expressed and crystallized in prelysed BMB171 cells (Fig. 2A and B). However, when almost half (at 40 h) or all (at 48 h) of the sporulation cell lysed, the free crystal of N586A was rarely observed (Fig. 2B and C). It seemed that the N586A crystal inclusion had been degraded or solubilized. Consistent with this, the crystal protein detection by SDS-PAGE also showed that the crystal proteins of N586A decreased at 40 h and became rare at 48 h (Fig. 2A).

Fig 2.

Protein expression and crystal formation of Cry5Ba and its mutant proteins. (A and B) Protein expression detection by SDS-PAGE (A) and crystal formation observation on a phase-contrast microscope (B) of Cry5Ba and its three mutant proteins in BMB171. (C) Sections through the crystal and spore of N586A and Cry5Ba, observed by TEM. (D and E) SDS-PAGE detection of crystal protein in culture precipitates (D) and soluble proteins in culture supernatants (E) of strain BMB171 producing N586A or Cry5Ba protein. The arrows in panels A, D, and E indicate the horizontal location of 140-kDa Cry proteins on a polyacrylamide gel. The arrows in panel B point to crystal inclusion in the mother cell or released from the mother cell.

To accurately validate the protein expression and crystal formation, the temporal expression of mutant protein N586A in strain BMB171 was further analyzed. As a result, similar to Cry5Ba, the mutant N586A protein was reliably expressed from 24 to 44 h (Fig. 2D). However, starting at 40 h, the amount of mutant N586A protein decreased in the culture precipitate (Fig. 2D) and increased in the culture supernatant (Fig. 2E). These observations suggested that the N586A crystal is easily dissolved once released from lysed sporulation cells.

The mutant N586A improved the crystal solubility across a wide range of pHs.

To test the hypothesis that N586A crystal might have an enhanced capacity to dissolve, the crystal solubility at different pHs was compared between N586A and Cry5Ba. It was found that the mechanically released N586A crystal is stable in 0.5 M NaCl–10 mM EDTA solution (pH 5.0) (Fig. 3A and B, lane Mix), as well as Cry5Ba. The collected crystals were allowed to incubate in carbonate buffer with different pHs for solubility analysis. In contrast to Cry5Ba, the N586A crystal exhibited considerable solubility in 50 mM Na₂CO₃ buffer at pH 5.0 to 7.0 (Fig. 3B), although both of them displayed equivalent abilities to dissolve at pH 8.0 to 10.0. The electron microscopic observation also showed that the N586A crystal, released from sporulation cell in 0.1 M PBS (pH 5.0, 6.0, and 7.0), displayed an appearance of partial dissolution different from that of Cry5Ba (Fig. 3C and D). These results suggest that the N586A crystal exhibits improved solubility at a wide range of pH values compared to Cry5Ba.

Fig 3.

Crystal solubility of N586A and Cry5Ba at different pHs. (A) Crystals of N586A and Cry5Ba in prelysed sporulation cell and artificially released from the mother cell at pH 5.0 by the ultrasonic process. Arrows point to the crystal inclusions. (B) Crystal solubility analysis of N586A compared to Cry5Ba in dissolution buffer with different pHs from 5.0 to 10.0. The solubilization supernatants were collected and loaded as SDS-PAGE samples. “Mix” indicates a mixture of spore and crystal after the ultrasonic process. An arrow indicates the horizontal location of 140-kDa Cry proteins on the polyacrylamide gel. (C and D) Crystals of N586A and Cry5Ba, released from a sporulation cell by incubation in 0.1 M PBS (pH 5.0, 6.0, and 7.0) at 4°C for 24 h, and observed by TEM (negative staining) (C) and SEM (D). All images in panels C and D were photographed at the same magnification of ×4,000.

The mutant N586A protein shows increased toxicity against the nematode C. elegans.

At the same time, to determine the effect of residue substitution on protein toxicity, the growth-inhibiting activity against nematode C. elegans larvae was qualitatively tested by using purified Cry5Ba and mutant proteins. The data showed that worms fed a higher dose of crystal proteins were paler in body (see Fig. S2 in the supplemental material) and smaller in size (Fig. 4B and Fig. S2) than those fed a lower dose of the proteins, suggesting that all three mutant proteins still retained their nematode toxicity. Furthermore, statistical analyses (Table 2) showed that the mutant N586A proteins had an increased toxicity, with a GIC₅₀ (that is, the medium growth-inhibiting concentration) of 4.75 ng/ml, nearly one-ninth that of Cry5Ba, whereas the other two mutant proteins N588A and N589A showed comparable toxicity to Cry5Ba.

Fig 4.

Toxicity assay on the nematode C. elegans with purified Cry5Ba and its mutant proteins. (A) SDS-PAGE detection of purified crystal proteins N586A, N588A, N589A, and Cry5Ba. (B) Three-day worm size of C. elegans fed different doses of purified crystal proteins in a growth assay. (C) SDS-PAGE detection of purified crystal protein N586A and Cry5Ba. (D) Three-day brood size of C. elegans fed with different doses of purified crystal proteins in the brood size assay. **, Differences were highly significant with a P < 0.01. In panels B and D, the data points represent the mean size relative to the control from five independent trials, and error bars denote the standard deviation from the mean.

Table 2.

Toxicity of purified crystal proteins (N586A, N588A, N589A, and Cry5Ba) against the nematode C. elegans

Toxin	GIC50a (ng/ml)	95% CIb (ng/ml)	Toxicity relative to Cry5Ba
N586A	4.75	1.58–14.29	8.87
N588A	29.16	12.36–68.80	1.44
N589A	33.01	13.86–78.64	1.28
Cry5Ba	42.11	18.37–97.87	1.00

^aGIC₅₀, median growth-inhibiting concentration.

^b95% CI, 95% confidence interval.

To give further evidence of enhanced toxicity of N586A protein, a brood size assay was further carried out on adult nematodes. The 3-day progeny number (Fig. 4D) showed that the mutant N586A toxin caused a smaller brood size than Cry5Ba. The statistical analysis indicated that the mutant N586A toxin was significantly different (P < 0.01) from Cry5Ba in incubation-inhibiting activity.

Mutant N586A toxin damages host intestine more efficiently than Cry5Ba.

Previous studies demonstrated that Cry5Ba toxin damages nematode intestine by shrinking the gut away from the body wall (30). To determine whether the mutant toxin N586A causes similar damage, the gut morphology of nematodes fed purified crystal proteins was attentively observed. The worm images in visible light (Fig. 5) showed that the mutant N586A and Cry5Ba proteins all induced shrinkage of the gut away from the body wall and resulted in constricted appearance. However, the mutant N586A toxin caused the gut to dwindle more rapidly compared to Cry5Ba when given at the same dose (Fig. 5).

Fig 5.

Intestinal damage assay on the nematode C. elegans with purified mutant N586A and Cry5Ba proteins at the same dose of 500 ng/ml. The worm intestine can emit blue autofluorescence upon excitation at a wavelength of 330 to 385 nm. The worm head is to the upside in each image; arrows indicate the edge of the worm body near the median of the gut.

Furthermore, we found that it can emit blue autofluorescence from the nematode intestine upon excitation at wavelengths of 330 to 385 nm, which facilitates the detection of intestine shrinkage induced by toxin. The extent and intensity of fluorescent signal diminishment in host intestine (Fig. 5) suggest that the mutant N586A toxin caused more significant damage of nematode intestine than did Cry5Ba, implying a more efficient action of N586A toxin than its native form.

DISCUSSION

The conserved blocks in B. thuringiensis Cry proteins were thought to be involved in conformational stability and interdomain contacts since these conserved blocks lie in the interface of domains or the core of individual domain (21). This study using site-directed mutagenesis partly emphasized the importance of Cry5Ba block 3 in protein solubility and toxicity, and the present study provides additional evidence for the involvement of conserved blocks in Cry protein's structural and functional role.

In the nematode intestine damage assay, we found that the intestine of a healthy C. elegans could automatically emit strong blue fluorescence, with an emission wavelength of 420 nm, upon excitation in a wavelength of 330 to 385 nm. We thought that this might facilitate the detection of intestinal change induced by B. thuringiensis toxins. As expected, we found that the toxin-treated worms fed Cry5Ba or N586A protein displayed reduced autofluorescence in the intestines and that the fluorescent signal diminishment was dependent on both incubation time (Fig. 5) and toxin dose (data not shown). Nematode autofluorescence was previously reported as an indicator of worm viability and survival (14) and for the detection of nematode eggs and other protozoa in swine feces (11). Coupled with our findings, these reports suggest that the nematode's autofluorescence might be a powerful tool in etiology and pathology research.

Crystal solubility has been previously demonstrated to be significant for their toxicity (1, 45). Moreover, the proteolytic activation of protoxin, a key step in toxicity mechanism of Cry proteins, was greatly dependent on crystal dissolution in host intestine (36). Some types of crystals from noninsecticidal B. thuringiensis isolates was soluble only at pH 12, and this lack of solubility in the host intestine (pH 9.0) contributed to their lack of toxicity (13). However, when presolubilized at a high pH, these nontoxic crystals did exhibit significant toxicity to their specific insect (39). In the present study, the enhanced toxicity of the mutant N586A protein might be partly due to its improved solubility especially at a low pH (5.0 to 7.0). Consistent with this, our pH assessment with different acid-base indicators (unpublished data) suggested that the pH value is ∼5.0 in the intestine of the nematode C. elegans. The pH optimum for most protease activity in intestinal crude extracts from C. elegans was found to be ∼5 (17, 42), implying the possibility of a mildly acidic pH in the intestinal lumen (31).

Structural flexibility is thought to be necessary for pore-forming proteins (38). According to the well-accepted mode of action, the activated toxins undergo a large conformational change to convert from the water-soluble to the membrane state (37), thereby demonstrating their structural flexibility. It was reported that Cry1A mutants in which a single salt bridge had been abolished exhibited increased molecular flexibility and a resultant faster pore formation (10). In the case of the Cry5Ba mutant N586A, the asparagine substitution probably led to a loss of polarity or the disruption of the hydrogen bond near the interface of the three domains, causing increased flexibility of the toxin and allowing conformational change for the insertion of the toxin into the membrane.

The fact that the Cry5Ba mutant N586A can be highly expressed (Fig. 2A and D) and form crystals (Fig. 2B and C) suggests that the mutant did not produce drastic alterations in the protein's structure. The similar CD spectra of the purified proteins N586A and Cry5Ba (see Fig. S3 in the supplemental material) also revealed that there was not much difference in their second structures. However, according to the protease digestion assay, the proteins N586A and Cry5Ba exhibited considerable differences in the profile of the degradation product, although they showed similar sensitivities to trypsin (see Fig. S4 in the supplemental material). The result implied that there may be a difference in the protein conformation or potential to convert in conformation between N586A and Cry5Ba.

In conclusion, we demonstrated that the asparagine substitution mutant N586A of Cry5Ba protein in block 3 improved the crystal solubility and toxin activity. We suggest that the amide residue removal at the interface of putative domains might facilitate toxin conformational transformation and membrane insertion. It is also clear that the residue substitution did not apparently affect the protein expression and crystal formation, suggesting that the mutant did not produce extreme structural changes. It may be practicable to use the mutant N586A gene to control pathogenic nematodes by plant genetic modification, since the mutant protein shows increased solubility and toxicity compared to the native toxin.