ABSTRACT
A novel recombinant Bacillus thuringiensis subsp. israelensis strain that produces the B. sphaericus binary toxin, Cyt1Aa, and Cry11Ba is described. The toxicity of this strain (50% lethal concentration [LC50] = 1.7 ng/ml) against fourth-instar Culex quinquefasciatus was higher than that of B. thuringiensis subsp. israelensis IPS-82 (LC50 = 7.9 ng/ml) or B. sphaericus 2362 (LC50 = 12.6 ng/ml).
Certain endotoxin proteins produced by Bacillus thuringiensis and Bacillus sphaericus are highly toxic to mosquito larvae (1, 4, 13). Examples are Cry4A, Cry4B, and Cry11A of B. thuringiensis subsp. israelensis and the binary (Bin) toxin of B. sphaericus 2362, which consists of a 51-kDa binding domain (BinA) and a 42-kDa toxin domain (BinB). Recently, a new mosquitocidal protein, Cry11B, from B. thuringiensis subsp. jegathesan was characterized (3). Cry11B is 58% identical to Cry11A but more toxic than the latter, the most toxic mosquitocidal protein produced by B. thuringiensis subsp. israelensis (2).
The development of highly effective recombinant B. thuringiensis subsp. israelensis strains with novel combinations of toxins is of considerable interest due to their potential utility in mosquito control. In a previous study, the toxicity of B. thuringiensis subsp. israelensis, which produces Cry4A, Cry4B, Cry11A, and Cyt1A (9), was improved considerably when cry11B was expressed in this strain using cyt1A promoters and the STAB-SD sequence (11, 12). Here we show that a recombinant B. thuringiensis strain that synthesized only the Bin toxin, Cyt1A and Cry11B, was significantly more toxic to Culex quinquefasciatus than either B. thuringiensis subsp. israelensis IPS-82 or B. sphaericus 2362.
Two plasmids (Fig. 1) were used for expression of toxin genes in B. thuringiensis subsp. israelensis 4Q7 (Bacillus Stock Center, Ohio State University, Columbus, Ohio). The first, p45S1, contained bin, cyt1A, and an erythromycin resistance gene (erm). To construct this plasmid, the 2.7-kb fragment containing the B. sphaericus 2362 bin operon was amplified by PCR with Vent (Exo+) DNA polymerase (New England Biolabs) using B. sphaericus 2362 genomic DNA. The PCR product was cloned into the filled XbaI and PstI sites in pSTAB-SD (12) containing a 660-bp cyt1Ap/STAB sequence. The cyt1Ap/STAB/bin fragment was then obtained by PCR using primers 5′-GGAATTCATTTTCGATTTC-3′ and 5′-AACTGCAGCCAAACAACAACAGTTTACATTCGAGTGTAAAAGTTC-3′ and cloned into the SmaI site of pWF45, whichcontains the cyt1A and 20-kDa-protein genes (16). The second plasmid, pPFT11Bs-CRP, contained cry11B and a chloramphenicol resistance gene (chl). To construct this plasmid, the 3-kb fragment containing the cyt1Ap/STAB/cry11B gene was amplified by PCR from pPFT11Bs (9) using primers 5′-ccgctcgagCGGGTCGACTATTTTCGATTTCAAATTTTCCAAACTT-3′ and 5′-ccgctcgagAAGCTTTTGTATGCCATCAAGAAAAAA-3′. This fragment was digested with XhoI and cloned in the same site in pHTC, a pUC19-based vector that contained the chl gene from pC194 (6) and the B. thuringiensis replication origin (ori) from pHT3101 (8), to generate pPFT11Bs-CRP (Fig. 1). The ori sequence was identical in p45S1 and pPFT11Bs-CRP.
The p45S1 or pPFT11Bs-CRP plasmid was introduced into B. thuringiensis subsp. israelensis 4Q7 by electroporation as described previously (10). Recombinant strains 4Q7/p45S1 and 4Q7/pPFT11Bs-CRP were selected on brain heart infusion agar containing, respectively, erythromycin (25 μg/ml) or chloramphenicol (10 μg/ml). The 4Q7/p45S1 strain was subsequently transformed with pPFT11Bs-CRP to generate 4Q7/p45S1-11B, which was resistant to erythromycin and chloramphenicol and produced Cyt1A, Cry11B, and the Bin toxin.
The 4Q7/p45S1, 4Q7/pPFT11Bs-CRP, and 4Q7/p45S1-11B strains were grown in 50 ml of nutrient broth plus glucose (NBG) (12) for 5 days at 30°C with antibiotic selection. During sporulation, 4Q7/p45S1-11B produced three distinct crystals (Fig. 2). After more than 95% of cells had lysed, 1 ml of each culture was pelleted, and the endotoxin proteins were separated for sodium dodecyl sulfate-polyacrylamide gel electrophoresis as described previously (7). The gel was stained with 0.125% Coomassie brilliant blue R-250, destained, and scanned by the GAS 4000 gel documentation system (Evergene). The amount of protein in each band was quantified with ImageQuant 4.1 densitometry software (Molecular Dynamics, Sunnyvale, Calif.) according to established methods (11).
Results showed a high level of endotoxin production in strains transformed with p45S1, pPFT11Bs-CRP, or with both plasmids (Fig. 2 and 3). However, the amount of Cyt1A produced by 4Q7/p45S1-11B decreased by 50% compared to that of 4Q7/p45S1, whereas no significant differences in yield were observed for Bin and Cry11B.
Lyophilized powders from NBG cultures were used to determine toxicity of the recombinants against fourth-instar larvae of C. quinquefasciatus and Aedes aegypti (kindly provided by M. C. Wirth, Department of Entomology, University of California, Riverside, Calif.). Controls used for comparison were B. thuringiensis subsp. israelensis IPS-82 (Institut Pasteur, Paris, France) and B. sphaericus 2362 (14, 15). Six to seven different concentrations of each powder were used to determine median lethal concentrations (LC50s), with 50 larvae for each concentration assayed. Mortality was recorded after 24 h of incubation at 28°C, and LC50s were calculated using Probit analysis (SAS Institute).
The strain 4Q7/p45S1-11B, which produced Cry11B, Cyt1A, and Bin, showed the highest toxicity against C. quinquefasciatus (Table 1), with an LC50 of 1.7 ng/ml. In addition to its high toxicity, an advantage of this strain is that the presence of Cyt1A may delay the development of resistance to Cry11B and the Bin toxin (5, 15), though this remains to be tested. The toxicity of the other 4Q7 recombinants was considerably lower. The strain that produced only Cry11B (4Q7/pPFT11Bs-CRP) had an LC50 of 9.2 ng/ml, whereas the strain that produced Cyt1A in combination with the Bin toxin (4Q7/p45S1) exhibited an LC50 of 3.7 ng/ml (Table 1). Against A. aegypti, however, toxicity of the recombinant that produced Cry11B, Cyt1A, and Bin (4Q7/p45S1-11B) was not significantly different from that of the strain that produced only Cry11B (4Q7/pPFT11Bs-CRP). Moreover, the Cyt1A plus Bin strain (4Q7/p45S1) was approximately ninefold less toxic to A. aegypti than the strain (4Q7/pPFT11Bs-CRP) that produced only Cry11B (Table 1). These results demonstrate that Cyt1A and Bin do not interact synergistically with Cry11B to improve its toxicity against A. aegypti.
The results reported here show that the combination of Cry11B, Cyt1A, and the Bin toxin is significantly more toxic to C. quinquefasciatus larvae, but not to those of A. aegypti, than the wild-type strain of B. thuringiensis subsp. israelensis or B. sphaericus 2362. These last strains serve as the active ingredients in commercial bacterial larvicides used for mosquito control. From a technical standpoint, our results also demonstrate that crystal toxin genes can be introduced into B. thuringiensis independently using plasmid vectors that harbor different antibiotic resistance markers. The identical ori sequence present in both plasmids we constructed could potentially affect plasmid stability, copy number, gene dosage, and endotoxin net production. However, as assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis, endotoxin production was consistent in sequential replicated growth trials under dual antibiotic selection (data not shown). Thus, this strategy could be useful for generating individual B. thuringiensis strains that produce various combinations of insecticidal proteins to assess their potential synergistic or antagonistic interactions.
Maps of recombinant plasmids and strategy for constructing a strain of B. thuringiensis that produces Cyt1A, Cry11B, and the B. sphaericus 2362 binary toxin. (A) p45S1 containing cyt1A from B. thuringiensis subsp. israelensis and a binary toxin gene from B. sphaericus 2362. (B) pPFT11Bs-CRP containing cry11B from B. thuringiensis subsp. jegathesan. Amp, ampicillin resistance gene; Erm, erythromycin resistance gene; Cm, chloramphenicol resistance gene; cyt1A-p, cyt1A promoters; cry1Ac-p, cry1Ac promoters; E. c. Ori, E. coli replication origin; B. t. Ori, B. thuringiensis replication origin.
Phase-contrast micrograph of B. thuringiensis subsp. israelensis strain 4Q7/p45S1-11B that produces crystals of Cry11B, Cyt1A, and the B. sphaericus binary toxin.
Analysis of endotoxin content in wild-type and recombinant strains of B. thuringiensis. M, molecular size marker; lane 1, B. thuringiensis subsp. israelensis 4Q7 producing B. sphaericus binary toxin and Cyt1A (4Q7/p45S1); lane 2, B. thuringiensis subsp. israelensis 4Q7 producing Cry11B (4Q7/pPFT11Bs-CRP); lane 3, B. thuringiensis subsp. israelensis 4Q7 producing Cry11B, Cyt1A, and B. sphaericus binary toxin (4Q7/p45S1-11B). The numbers at the base of lane 3 indicate the approximate ratio of each toxin produced in the Cry11B, Cyt1A, Bin recombinant in comparison to, respectively, the Cyt1A plus Bin recombinant (lane 1) and the Cry11B recombinant (lane 2).
Toxicity of B. thuringiensis subsp. israelensis 4Q7 strains producing the B. sphaericus binary toxin, B. thuringiensis Cry11B, and/or Cyt1A against fourth-instar C. quinquefasciatus and A. aegypti
ACKNOWLEDGMENTS
We thank Jeffrey J. Johnson for his technical assistance during this study.
This research was supported in part by grants from the United States National Institutes of Health (AI45817), the University of California BioSTAR program (99-10070), and the University of California Mosquito Research Program.
FOOTNOTES
- Received 19 August 2002.
- Accepted 5 November 2002.
- American Society for Microbiology
