Introduction of Culex toxicity into Bacillus thuringiensis Cry4Ba by protein engineering

Abstract

Bacillus thuringiensis mosquitocidal toxin Cry4Ba has no significant natural activity against Culex quinquefasciatus or Culex pipiens (50% lethal concentrations [LC(50)], >80,000 and >20,000 ng/ml, respectively). We introduced amino acid substitutions in three putative loops of domain II of Cry4Ba. The mutant proteins were tested on four different species of mosquitoes, Aedes aegypti, Anopheles quadrimaculatus, C. quinquefasciatus, and C. pipiens. Putative loop 1 and 2 exchanges eliminated activity towards A. aegypti and A. quadrimaculatus. Mutations in a putative loop 3 resulted in a final increase in toxicity of >700-fold and >285-fold against C. quinquefasciatus (LC(50) congruent with 114 ng/ml) and C. pipiens (LC(50) 37 ng/ml), respectively. The enhanced protein (mutein) has very little negative effect on the activity against Anopheles or AEDES: These results suggest that the introduction of short variable sequences of the loop regions from one toxin into another might provide a general rational design approach to enhancing B. thuringiensis Cry toxins.

FIG. 1.

Sequence alignments based on the model structures made with Swiss-Pdb Viewer (18) of Cry4Aa with Cry4Ba. Loop positions are indicated above the sequences, while the amino acid residues involved are in boldface.

FIG.2.

CD spectra of purified toxins of Cry4Ba and its mutants.

FIG.3.

Specific saturation binding of 4BL3PAT and 4BRA to C. quinquefasciatus BBMV.

FIG. 4.

(A) Homologous and heterologous competition binding assays. ¹²⁵I-labeled 4BRA was incubated with C. quinquefasciatus BBMV with increasing amounts of unlabeled toxin. (B) Irreversible binding studies. C. quinquefasciatus BBMV were preincubated with ¹²⁵I-labeled toxin for 1 h. Then binding was competed with excess cold toxin for different durations. Data shown are means of three binding experiments.

FIG. 5.

Proteinase K protection assay of 4BRA and 4BL3PAT. ¹²⁵I-labeled toxin was incubated with C. quinquefasciatus BBMV for 1 h. Free and noninserted toxin was digested with proteinase K, and the reaction was stopped with Pefabloc. BBMV-protected toxin was separated by centrifugation, and the counts were measured.

FIG. 6.

C. quinquefasciatus BBMV light-scattering assay. BBMV (0.2 mg/ml) were coinjected with different samples. All samples were in 10 mM HEPES buffer, pH 7.5. The samples were KCl (150 mM) (line 1), 4BRA (66 pmol of toxin/mg in 150 mM KCl) (line 2), 4BL3PAT (66 pmol of toxin/mg in 150 mM KCl) (line 3), and HEPES buffer alone (line 4). Light scattering was measured using a stopped-flow apparatus at a 488-nm wavelength with the temperature controlled at 20°C. The curves were normalized to begin from the same point. The increasing intensity is indicative of reswelling of the vesicles. The data shown are averages of two experiments.

FIG. 7.

Ligand blot of biotin-labeled 4BRA (Cry4Ba) and 4BL3PAT (Mutant) to C. quinquefasciatus BBMV. Ten micrograms of BBMV proteins was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, blotted onto a polyvinylidene difluoride membrane, and probed with 5 nM biotinylated toxins. Arrows point to bands that are uniquely bound by 4BL3PAT. Lanes 1 and 3, BBMV; lanes 2 and 4, precipitated BBMV. Numbers between the panels are molecular masses in kilodaltons.

FIG. 8.

Model for binding of toxic 4BL3PAT to productive binding sites (gray) and nonproductive binding sites (white). Nontoxic protein 4BRA binds only to the nonproductive binding sites. The nonproductive binding sites are more numerous.