Alanine-substituted mutant on Gly373 and Asn375 of Cry1Ai-h-loop 2 causes reduction in both toxicity and binding against Helicoverpa armigera

LIU Yu-xiao, ZHOU Zi-shan, LIANG Ge-mei, SONG Fu-ping, ZHANG Jie

State Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing 100193, P.R.China

Abstract Cry1Ai-h-loop 2 is a mutant of Cry1Ai constructed by exchanging loop 2 from Cry1Ah protein and shows insecticidal activity against Helicoverpa armigera. The toxicity of Cry1Ai-h-loop 2, in contrast to the very low toxicity of Cry1Ai, is closely associated with the eleven residues in the loop 2 region. To characterize the key sites of loop 2 in Cry1Ai-h-loop 2, alaninesubstituted mutants were generated. The toxicity of these mutants against H. armigera indicated that dual-mutant on Gly373 and Asn375 caused a signif icant decrease in toxic activity. ELISA binding and competition binding assays demonstrated that the reduction of toxicity in the mutant of interest was correlated with decreased binding aff inity.

Keywords: Bacillus thuringiensis, Cry1Ai, Domain II-loop2, Helicoverpa armigera, binding aff inity

1. Introduction

Bacillus thuringiensis (Bt), a widely used and successful bioinsecticide, is a spore-forming and Gram-positive microbe with specif ic toxicity against diverse insect orders such as Lepidoptera, Diptera, and Coleoptera, as well as against nematodes and human cancer cells (Palma et al. 2014; Lacey et al. 2015; Jouzani et al. 2017). Cry toxins are biologically active parasporal crystal proteins produced during the growth cycle of Bt (Schnepf et al. 1998). To date over 790 cry genes have been identif ied and named (Crickmore et al. 2016). Despite low amino acid sequence similarity, the structural similarity between different trypsinactivated toxins that attach to three core domains Cry (3d-Cry) toxins is substantially high (Pardo-López et al. 2013). It is generally accepted that the three domains of the toxin play different roles: Domain I, composed of seven α-helices, is involved in oligomerization, membrane insertion, and pore formation; Domain II, formed by β-sheets, is implicated in receptor interactions; and Domain III, comprising a β-sandwich structure, participates in receptor binding (Bravo et al. 2007).

The toxic mechanism of Cry1A against target Lepidoptera has been widely investigated. In sequential binding model, it was established that for the toxicity to be exerted, the activated toxins must sequentially bind with specif ic receptors on the surface of brush border membrane vesicles (BBMV) of the insect midgut, thereby causing membrane insertion and pore formation (Bravo et al. 2004; Pigott and Ellar 2007; Pacheco et al. 2009; Pardo-López et al. 2013). Additionally, an alternative model of mechanism was involved in binding of Cry1Ab with receptor BT-R1, then provoking cell death by signaling pathway (Zhang et al. 2006). All research on the mechanism indicates that binding with BBMV is a crucial step in the toxicity of Cry1A toxins.

Cry1Ai and Cry1Ah are encoded by cry1Ai2 and cry1Ah1, respectively. These genes were cloned by Institute of Plant Protection, Chinese Academy of Agricultural Sciences and the encoded proteins were toxic to many lepidopteran pests (Xue et al. 2008; Shu et al. 2013). Interestingly, even though the two proteins exhibit 84.72% of amino acid sequence identity, Cry1Ah exhibited strong insecticidal activity against Helicoverpa armigera, while Cry1Ai only induced weight loss against H. armigera (Xue et al. 2008; Shu et al. 2013). The loop-exchanged mutant Cry1Ai-h-loop 2, containing loop 2 from Cry1Ah, exhibited signif icantly higher toxicity than that of Cry1Ai, while the mutant Cry1Ai-h-loop 3, containing loop 3 from Cry1Ah, was unchanged in terms of activity (Zhou et al. 2017). This f inding suggested that the loop 2 residues (368RPFNIGINNQQ378) in the Cry1Ai-h-loop 2 mutant determine the specif icity and toxicity against H. armigera. The key sites within this loop, however, remain unclear.

Alanine site-directed mutagenesis is a common approach for characterizing protein functional sites and has been used in studies on Cry protein toxicity (Khaokhiew et al. 2009; Lucena et al. 2014), receptor binding (Gómez et al. 2014), pore formation (Juntadech et al. 2014), solubility (Wang et al. 2012), and synergism (Pérez et al. 2005; Cantón et al. 2011). Therefore, in this study, alanine-substituted mutants on loop 2 were constructed for further characterization and identif ication of the functional sites in Domain II loop 2 region of Cry1Ai-h-loop 2 against H. armigera.

2. Materials and methods

2.1. Sequence and structure analyses

The 3D-structure of Cry1Ai-h-loop 2 was predicted on SWISS-MODEL workspace (https://www.swissmodel.expasy.org) and analyzed using PyMOL v.0.99rc6 Software (http://www.pymol.org). The DNA and protein sequences were analyzed using DNAMAN v.6.0.3.48 analysis tools (http://www.lynnon.com).

2.2. Construction of alanine-substituted mutants

The Cry1Ai-h-loop 2 plasmid constructed in our previous study (Zhou et al. 2017) was used as a template for amplifying the mutations using the Trans?Fast Mutagenesis System (TransGen, Beijing, China). The primers listed in Table 1 were designed according to the manufacturer's instructions. The resulting PCR products were digested by DMT enzyme to degrade the methylated template plasmid and were then used to transform DMT competent cells.

The success of the mutation step was conf irmed by DNA sequencing.

2.3. Expression of Cry1Ai-h-loop 2 and alanine-sub-stituted mutants

The constructed plasmids were used to transform Escherichia coli Rosetta DE3 cells (Novagen, Madison, WI). Competent transformants were precultured in 5 mL Luria-Bertani medium containing 100 μg mL-1ampicillin and 34 μg mL-1chloromycetin for 8 h and were then transferred into 400 mL LB medium (100 μg mL-1ampicillin and 34 μg mL-1chloromycetin). The cells were cultured at 37°C with shaking at 220 r min-1, and at an optical density (OD600) of 0.6, IPTG was added to medium with 0.5 mmol L-1(f inal concentration) to induce the protein expression for 12 h at 20°C, and 150 r min-1. Cells were collected by centrifugation at 10 000×g for 5 min and the resulting pellets were resuspended using 20 mmol L-1Tris-HCl buffer (p H 8.0) before being sonicated (Ningbo Scientz Biotechnology Co., Ltd., China) for 5 min (3 s on and 5 s off) at 75% power. Inclusions were isolated by centrifugation at 12 000×g for 10 min at 4°C, after which the resulting supernatants were transferred and the pellets were resuspended in 20 mmol L-1Tris-HCl buffer (p H 8.0). The soluble materials and insoluble precipitates were analyzed by 8% SDS-PAGE. The quantif ication of insoluble protein was determined as described previously (Zhou et al. 2017). The optical density of interest binds stained by Coomassie blue was calculated by Image J Software with bovine serum albumin (BSA) as standard. The alanine-substituted mutants were also expressed and isolated as described here.

2.4. Insect bioassay

The toxicity of Cry1Ai-h-loop 2 and alanine-substituted mutants against H. armigera neonatal larvae was assessed by mixing resuspended pellets (described in Section 2.3) with an artif icial diet (Wu et al. 1999). Each protein was assayed in triplicate. Twenty-four insects were individually inoculated into 24-well plates containing 400 mg diet per well and mortality was calculated after 7 d.

2.5. Preparation, purif ication, and biotinylation of activated toxins

Cry1Ai-h-loop 2 and dual-mutant G373A and N375A were truncated from Met1to Leu636, which fragment includes the minimal active fragment of Cry1Ai (Zhou et al. 2014). Firstly, Cry1Ai was truncated from Met1to Leu636(T-Cry1Ai) using the primers listed in Table 1 as described previously (Zhou et al. 2014). Next, truncated Cry1Ai-h-loop 2 (T-Cry1Aih-loop 2) and dual-mutant (T-G373A and N375A) were amplif ied by reverse PCR using the primers listed in Table 1 with T-Cry1Ai (Met1to Leu636) as a template as described previously (Zhou et al. 2017). The truncated proteins were expressed as described above and the soluble protein was purif ied by nickel aff inity chromatography using Chelating Sepharose? Fast Flow media (GE Healthcare, Uppsala, Sweden). The soluble component isolated by centrifugation after sonication was f lowed through media pre-charged with Ni2+, after which nonspecif ically binding protein was eluted using buffer containing 20 mmol L-1Tris-HCl, 500 mmol L-1NaCl, and 100 mmol L-1imidazole (p H 8.0). The target protein was f inally eluted with buffer containing 20 mmol L-1Tris-HCl, 500 mmol L-1NaCl, and 250 mmol L-1imidazole (p H 8.0). Toxins were activated by trypsin and labeled using EZ-Link Sulfo-NHS-SS-Biotin (Thermo Fisher Scientif ic, Rockford, IL, USA) as per the manufacturer's instructions. Following the biotinylation step carried out at room temperature, size-exclusion chromatography (HiTrap? Desalting, GE Healthcare, Uppsala, Sweden) was used to remove excess free biotin.

2.6. Preparation of BBMV from H. armigera

H. armigera larvae, kindly supplied by Cotton Insect Pests Laboratory, Institute of Plant Protection, Chinese Academy of Agricultural Sciences, were reared on an artif icial diet.

Third instar larvae were dissected and the midgut tissue was exteriorized after food debris and fat bodies were removed. The tissue was then washed with PBS buffer and stored at -70°C. The BBMV were prepared as described by Wolfersberger et al. (1987).

2.7. ELISA binding and competition binding assays

The wells of a polystyrene 96-well microplate (Costar, Kennebunk, ME) were coated with 1 μg H. armigera BBMV at 4°C overnight. After washing three times using PBS, the wells were blocked with PBS containing 2% BSA at 37°C for 2 h before being incubated with increasing amounts of biotinylated, activated Cry1Ai-h-loop 2 and dual-mutant G373A and N375A toxin at 37°C for 1 h. After washing three times with PBST (PBS containing 0.1% Tween-20), Pierce? High Sensitivity Streptavidin-HRP (Thermo Fisher Scientif ic, Rockford, IL; 1:40 000) was placed into the wells at 37°C for 1 h before being washed with PBST for three times. Finally, 100 μL 3,3′,5,5′-tetramethylbenzidine (TMB) substrate solution (Solarbio, Beijing, China) was ad ded and absorbance was measured at 450 nm.

Table 1 Primers used for mutagenesis

For competition binding assays, two peptides from loop 2, wild type peptide368RPFNIGINNQQ378on Cry1Ai-h-loop 2, and mutational peptide368RPFNIAIANQQ378on dual-mutant G373A and N375A, were synthesized (Sangon, Shanghai, China) as competitors. Wells coated with H. armigera BBMV and blocked with 2% BSA were incubated with premixed biotinylated Cry1Ai-h-loop 2 toxin (30 nmol L-1) and synthetic wild type and mutant peptides (15 000 nmol L-1) at 37°C for 1 h before being incubated with Streptavidin-HRP and substrate solution as described above.

3. Results

3.1. Recombinant expression and toxicity of Cry1Aih-loop 2 and alanine mutants against H. armigera

To determine the key sites on Cry1Ai-h-loop 2 contributing to alterations in toxicity, ten amino acids from Pro369to Gln378in the loop 2 (Fig. 1) were substituted with alanine, since residues Arg368was already the same as in Cry1Ai. As shown in Fig. 2-A, all the mutants were expressed in E. coli and produced proteins with molecular mass of 130 kDa, which was similar to that of Cry1Ai-h-loop 2. As shown in Fig. 2-B, under the concentration of 100 μg g-1Cry1Ai-h-loop 2 resulted in (83±6.0)% corrected mortality of H. armigera. Most mutants exhibited similar toxicity to this, except for mutants G373A and N375A, which yielded signif icantly lower toxicity (P＜0.01), only inducing (43±10.4)% and (44±6.0)% corrected mortality, respectively. The dual-mutant G373A and N375A produced a protein with a similar molecular mass to that of Cry1Ai-h-loop 2. The dual-mutant did not exhibit major modif ications in structure, since the mutant was activated as Cry1Ai-h-loop 2 (Fig. 2-C); however, the toxicity of the dual-mutant G373A and N375A was decreased to (23±2.2)% mortality (compared with (76±6.0)% mortality due to Cry1Ai-h-loop 2) at 100 μg g-1(Fig. 2-D), suggesting that Gly373and Asn375were important in loop 2 region of Cry1Ai-h-loop 2.

Fig. 1 Homology-modeled 3D protein structure of Cry1Aih-loop 2 by SWISS-MODEL workspace using Cry1Ac toxin structure (PDB ID: 4ary) as a template. Analysis was carried out using PyMOL Software. The exposed loop 2 region (blue) is annotated and magnif ied on the left, where the positions of Gly373 and Asn375 on the loop are indicated by arrows.

3.2. Competition binding analysis of Cry1Ai-h-loop 2 and dual-mutant G373A and N375A with H. armigera BBMV via synthesized peptides

To determine whether the signif icantly decreased activity of the dual-mutant G373A and N375A is related to loss of binding, competition binding assays were performed using synthesized peptides to bind with BBMV and competed the binding between toxins and BBMV. The wild type peptide368RPFNIGINNQQ378from Cry1Ai-h-loop 2 and the mutational peptide368RPFNIAIANQQ378from dual-mutant G373A and N375A were added as competitors upon incubation with biotinylated toxin. As shown in Fig. 3-B, both peptides signif icantly competed with the binding of Cry1Ai-h-loop 2 to BBMV (P＜0.01); however, the mutant peptide368RPFNIAIANQQ378from the dual-mutant G373A and N375A was found to compete signif icantly less than the wild type peptide368RPFNIAIANQQ378(P＜0.01). These data indicated that the dual-mutant G373A and N375A, which exhibited decreased toxicity, did not lose the ability to bind to BBMV completely, but exhibited a diminished binding aff inity with H. armigera BBMV compared with that exhibited by Cry1Ai-h-loop 2.

3.3. Binding affnity of Cry1Ai-h-loop 2 and dual mutant G373A and N375A with H. armigera BBMV

The binding aff inities of Cry1Ai-h-loop 2 and the dualmutant G373A and N375A with BBMV from H. armigera were measured by ELISA. First, truncated proteins were purif ied, activated, and biotinylated (Fig. 4-A), after which increasing molar concentrations of biotinylated toxin were allowed to bind with 1 μg of BBMV. The resulting saturation binding curves were shown in Fig. 4-B. Cry1Ai-h-loop 2 exhibited a greater (almost two-fold) BBMV binding aff inity (Kdof (24.54±2.97) nmol L-1) than the dual-mutant G373A and N375A (Kdof (49.53±15.94) nmol L-1) at low levels.

4. Discussion

Alanine mutagenesis is a frequently used technique for characterizing protein functional sites (Khaokhiew et al. 2009; Wang et al. 2012; Gómez et al. 2014; Juntadech et al. 2014) and mutating functional residues in Cry proteins normally led to variations to toxicity (Tigue et al. 2001; Pacheco et al. 2009; Sengupta et al. 2013).

Fig. 2 Expression and insecticidal activity of Cry1Ai-h-loop 2 and mutants. A, the expression analysis of Cry1Ai-h-loop 2 and alanine mutants by using 8% polyacrylamide gel. The arrows indicated expected sizes of target proteins. B, the expression analysis of soluble Cry1Ai-h-loop 2, dual-mutant G373A and N375A and their respective trypsin-activated fragments by using 8% polyacrylamide gel and indicated by arrows. C, bioassay results of Cry1Ai-h-loop 2 and alanine-substituted mutants against Helicoverpa armigera neonatal larvae under the concentration of 100 μg g-1 of Cry1Ai-h-loop 2 and its mutants. Data represent means of three repeats and error bars indicate standard deviation of the means. Asterisks indicate statistically signif icant difference in mortality between Cry1Ai-h-loop 2 and mutant G373A or N375A (P＜0.01). D, bioassay results of Cry1Ai-h-loop 2 and dualmutant G373A and N375A against H. armigera neonatal larvae under the concentration of 100 μg g-1 of Cry1Ai-h-loop 2 and two dual-mutants. Data represent means of three repeats and error bars indicate standard deviation of the means. Asterisks indicate statistically signif icant difference in mortality between Cry1Ai-h-loop 2 and dual-mutant G373A and N375A (P＜0.01).

Loops 2 and 3 from Cry1Ah protein have previously been shown to be important for toxicity and binding to H. armigera (Zhou et al. 2017). Since a Cry1Ai-h-loop 2 mutant containing loop 2 (RPFNIGINNQQ) from Cry1Ah showed signif icant toxicity against H. armigera, alanine-substituted mutants were constructed based on Cry1Ai-h-loop 2 to identify the important residues in this region. The bioassay results showed that the toxicity of two mutants (G373A and N375A) was reduced, indicating that both Gly373and Asn375on loop 2 are functional sites. This association was further proven by the reduced toxicity of the dual-mutant G373A and N375A.

In Cry1A toxins, binding with BBMV is a crucial step in the toxicology. Loss or reduction of toxicity for these toxins has most often been shown to be due to decreased binding to BBMV or receptors (Lu et al. 1994; Rajamohan et al. 1996b; Atsumi et al. 2008; Roh et al. 2009); however, in some cases, binding and toxicity were shown to be unrelated (Howlader et al. 2010; Adegawa et al. 2017). In this case, the dual-mutant G373A and N375A only exhibited a decrease in binding aff inity for BBMV from H. armigera, not a complete loss of binding. This may because both loops 2 and 3 contributed to receptor binding (Zhou et al. 2017). The dual-mutant G373A and N375A furthermore still exhibited weight loss activity against H. armigera (data not shown) and may therefore still bound to some receptors.

Fig. 3 Synthetic peptides competing with Cry1Ai-h-loop 2 for binding to Helicoverpa armigera brush border membrane vesicles (BBMV). A, the sequences of synthetic peptides (framed) belonging to loop 2 of Cry1Ai-h-loop 2 and dual-mutant G373A and N375A. B, competition binding assay of Cry1Ai-hloop 2 and H. armigera via synthetic peptides as competitors and analyzed by ELISA. The columns show the means of absorbance values (n=3) and error bars denote standard deviation of the means. Different letters represent a signif icant difference at P＜0.01.

By comparing the 3D-structure of loop 2 region among Cry1Ai-h-loop 2, Cry1Ah, Cry1Ac and Cry1Ai toxins, as shown in Fig. 5, we found that the structure of loop 2 on Cry1Ai-h-loop 2, Cry1Ah and Cry1Ac was the same, while it was different with that of Cry1Ai. Gly and Asn were existed in loop 2 of all Cry1Ai-h-loop 2, Cry1Ah, Cry1Ac and Cry1Ai toxins, however, the structural conf iguration of these two residues was different (Fig. 5), which was mainly related with the insertion of Gly, two sites ahead from Gly and Asn, and the different amino acid residues on other sites of loop 2. Besides, the difference on the structure of sites Gly and Asn, as well as the whole loop 2 region, may have been a key factor towards the altering of toxicity between Cry1Ai and Cry1Ai-h-loop 2.

It has already been demonstrated in several studies that modif ications in the Domain II loops region of the toxin inf luence interactions with BBMV and eventually alter toxicity (Rajamohan et al. 1996a; Atsumi et al. 2008; Arenas et al. 2010). In contrast, research into the functional sites for toxicity towards H. armigera has been primarily focused on domain III (Wang et al. 2009; Liu et al. 2010; Lv et al. 2011; Sengupta et al. 2013). In this study, we successfully identif ied dual-mutant on Gly373and Asn375caused a signif icant decrease in toxicity and binding of Cry1Ai-h-loop 2 against H. armigera. These f indings taken together with previously reported f indings indicate that the exposed loops are essential for both toxicity and interaction. This study furthermore demonstrates that site-directed mutagenesis is an effective approach to determining the location and function of key residues.

5. Conclusion

In this study, Ala-substituted mutant on Gly373and Asn375reduces the toxicity and binding of Cry1Ai-h-loop 2 against H. armigera. Gly373and Asn375are important residues on the loop 2 region of Cry1Ai-h-loop 2.

Fig. 4 Cry1Ai-h-loop 2 and dual-mutant G373A and N375A bound with Helicoverpa armigera brush border membrane vesicles (BBMV). A, SDS-PAGE analysis of Cry1Ai-h-loop 2 and dual-mutant G373A and N375A by using 10% polyacrylamide gel after biotinylation. B, binding assay between biotinylated Cry1Ai-h-loop 2 and dual-mutant G373A and N375A with H. armigera BBMV by ELISA. Optical density was measured at 450 nm and the resulting saturation binding curves were plotted using SigmaPlot ver. 12.5 Software (Systat Software, CA). Error bars denote standard deviation of the means (n=3).

Fig. 5 Structural analysis of loop 2 region of Cry1Ai-h-loop 2, Cry1Ah, Cry1Ac and Cry1Ai. The 3D-structure of Cry1Ai-hloop 2, Cry1Ah was modeled by SWISS-MODEL workspace using Cry1Ac toxin structure (PDB ID:4ary) as a template, while the homology-model of Cry1Ai was builded based on Cry1Aa toxin structure (PDB ID:1ciy). Analysis was carried out using PyMOL Software. The exposed loop 2 region (blue) is from Cry1Ai-h-loop 2, Cry1Ah and Cry1Ac, while the pink one is from Cry1Ai. The positions of Gly373 and Asn375 on the loop are indicated by arrows.

Acknowledgements

We thank Dr. Neil Crickmore from School of Life Sciences, University of Sussex, Prof. Alejandra Bravo and Prof. Mario Soberón from Institute of Biotechnology, National Autonomous University of Mexico, and Mr. Wang Zeyu form Institute of Plant Protection, Chinese Academy of Agricultural Sciences for their assistance in this study. This work was supported by the National Key R&D Program of China (2017YFD0200400) and the National Natural Science Foundation of China (31272115).