Analysis of mosquito larvicidal potential exhibited by vegetative cells of Bacillus thuringiensis subsp. israelensis

Vegetative Bacillus thuringiensis subsp. israelensis cells (6 X 10(5)/ml) achieved 100% mortality of Aedes aegypti larvae within 24 h. This larvicidal potential was localized within the cells; the cell-free supernatants did not kill mosquito larvae. However, they did contain a heat-labile hemolysin which was immunologically distinct from the general cytolytic (hemolytic) factor released during solubilization of B. thuringiensis subsp. israelensis crystals. The larvicidal potential of the vegetative cells was not due to poly-beta-hydroxybutyrate. Instead, it correlated with the ability of vegetative cells to sporulate during the bioassays. No toxicity was observed when bioassays were conducted in the presence of chloramphenicol or streptomycin. It is unlikely that the vegetative cells sporulate in the alkaline (pH 9.5 to 10.5) larval guts after ingestion. B. thuringiensis subsp. israelensis is not an alkalophile; we have been unable to grow it in culture at pH values of greater than or equal to 9.5. Moreover, we have been unable to demonstrate formation of a protective capsule. However, bacteria may replicate in the gut fluids of dead or dying mosquito larvae because their alkaline gut pH values drop markedly after exposure to the B. thuringiensis subsp. israelensis crystal toxins.     

Variations in the mosquito larvicidal activities of toxins from Bacillus thuringiensis ssp. israelensis

Comparing activities of purified toxins from Bacillus thuringiensis ssp. israelensis against larvae of seven mosquito species (vectors of tropical diseases) that belong to three genera, gleaned from the literature, disclosed highly significant variations in the levels of LC₅₀ as well as in the hierarchy of susceptibilities. Similar toxicity comparisons were performed between nine transgenic Gram-negative species, four of which are cyanobacterial, expressing various combinations of cry genes, cyt1Aa and p20 , against larvae of four mosquito species as potential agents for biological control. Reasons for inconsistencies are listed and discussed. Standard conditions for toxin isolation and presentation to larvae are sought. A set of lyophilized powders prepared identically from six Escherichia coli clones expressing combinations of four genes displayed toxicities against larvae of three mosquito species, with levels that differed between them but with identical hierarchy.     

Comparison of development of Bacillus thuringiensis subsp. israelensis and Bacillus sphaericus in mosquito larvae

Immunofluorescent staining was used with thin sections of paraffin-embedded specimens to detect the development of Bacillus thuringiensis var. israelensis and Bacillus sphaericus in the gut of mosquito larvae. The third- and fourth-instar larvae of Aedes aegypti, Anopheles maculatus, and Culex quinquefasciatus were fed either vegetative cells or spores of the bacteria. Spore germination, multiplication, and sporulation were studied in the larvae of each species. The spores of B. thuringiensis var. israelensis and B. sphaericus strain 2297 could germinate and cells could sporulate in the larval body. The vegetative cells of B. sphaericus strain 810428 were also able to produce spores in the mosquito larval gut, but the germination of spores could not be detected in the larvae. Multiplication of all bacterial species was observed after the larvae died. Growth of the bacteria in distilled water containing crude extracts of larvae made from each species was compared with that in synthetic medium (nutrient broth). They could produce spores and toxins in all the media used and the toxins had larvicidal activity against the target mosquitos Ae. aegypti, An. maculatus, and C. quinquefasciatus.     

Micro-lipid-droplet encapsulation of Bacillus thuringiensis subsp. israelensis delta-endotoxin for control of mosquito larvae

The crystal delta-endotoxin of Bacillus thuringiensis subsp. israelensis is less toxic to larvae of Anopheles freeborni than to larvae of Aedes aegypti. However, when solubilized crystal was used, larvae from both species showed similar sensitivities. This effect presumably was due to the differences in feeding behavior between the two mosquito larvae when crystal preparations are used. A procedure is described whereby both crystal and solubilized B. thuringiensis subsp. israelensis toxin were emulsified with Freund incomplete adjuvant, with retention of toxicity. The use of Freund incomplete adjuvant also allowed one to assay the solubilized toxin at a low nanogram level. Furthermore, coating the toxin with lipophilic material altered the buoyancy of the toxin and reversed the sensitivities of the two mosquito larvae toward the B. thuringiensis subsp. israelensis toxin. This difference in buoyancy was determined by using an enzyme-linked immunosorbent assay that was specific for the toxic peptides. These data indicate that economically feasible buoyant formulations for the B. thuringiensis subsp. israelensis crystal can be developed.     

Micro-lipid-droplet encapsulation of Bacillus thuringiensis subsp. israelensis delta-endotoxin for control of mosquito larvae.

The crystal delta-endotoxin of Bacillus thuringiensis subsp. israelensis is less toxic to larvae of Anopheles freeborni than to larvae of Aedes aegypti. However, when solubilized crystal was used, larvae from both species showed similar sensitivities. This effect presumably was due to the differences in feeding behavior between the two mosquito larvae when crystal preparations are used. A procedure is described whereby both crystal and solubilized B. thuringiensis subsp. israelensis toxin were emulsified with Freund incomplete adjuvant, with retention of toxicity. The use of Freund incomplete adjuvant also allowed one to assay the solubilized toxin at a low nanogram level. Furthermore, coating the toxin with lipophilic material altered the buoyancy of the toxin and reversed the sensitivities of the two mosquito larvae toward the B. thuringiensis subsp. israelensis toxin. This difference in buoyancy was determined by using an enzyme-linked immunosorbent assay that was specific for the toxic peptides. These data indicate that economically feasible buoyant formulations for the B. thuringiensis subsp. israelensis crystal can be developed.     

Mosquito larvicidal activity of Escherichia coli with combinations of genes from Bacillus thuringiensis subsp. israelensis.

The genes cryIVA and cryIVD, encoding 134- and 72-kDa proteins, respectively, and the gene for a regulatory 20-kDa polypeptide of Bacillus thuringiensis subsp. israelensis (serovar H14) were cloned in all seven possible combinations by the Escherichia coli expression vectors pT7 and pUHE. The four combinations containing cryIVA (cryIVA alone, with cryIVD, with the 20-kDa-protein gene, and with both) displayed high levels of mosquito larvicidal activity in pUHE. The toxicity of the combination of cryIVA and cryIVD, with or without the 20-kDa-protein gene, was higher than has ever been achieved with delta-endotoxin genes in recombinant E. coli. Fifty percent lethal concentrations against third-instar Aedes aegypti larvae for these clones decreased (i.e., toxicity increased) continuously to about 3 x 10(5) cells ml-1 after 4 h of induction. Larvicidal activities, obtained after 30 min of induction, were lower for clones in pT7 and decreased for an additional 3.5 h. Induction of either cryIVD or the 20-kDa-protein gene alone resulted in no larvicidal activity in either pT7 or pUHE20. Cloned together, these genes were slightly toxic in pT7 but not in pUHE20. Five minutes of induction of this combination (cryIVD with the 20-kDa-protein gene) in pT7 yielded a maximal mortality of about 40%, which decreased rapidly and disappeared completely after 50 min. CryIVD is thus apparently degraded in E. coli and partially stabilized by the 20-kDa regulatory protein. Larvicidal activity of the combination of cryIVA and cryIVD was sevenfold higher than that of cryIVA alone, probably because of the cross-stabilization of the polypeptides or the synergism between their activities.     

Delta endotoxin of Bacillus thuringiensis subsp. israelensis.

From Bacillus thuringiensis subsp. israelensis, a proteinase-resistant protein was purified which exhibited toxicity to larval mosquitoes and cultured mosquito cells, lysed erythrocytes, and was lethal to mice. To extract the protein, a sporulating culture of B. thuringiensis subsp. israelensis was treated with alkali, neutralized, and incubated with trypsin and proteinase K. It was then purified by gel filtration and DEAE column chromatography. Up to 240 micrograms of toxic protein was purified from 1 g (wet weight) of culture pellet. Two closely related forms of toxic protein were obtained: the 25a and 25b proteins. The two forms comigrated near 25,000 daltons in a sodium dodecyl sulfate-polyacrylamide gel, were serologically related, and showed similar partial protease digestion profiles, but were distinguishable by DEAE chromatography and nondenaturing polyacrylamide gel electrophoresis. Protein sequencing data indicated the 25b protein lacked the two amino acids at the amino terminus of the 25a protein. A Western blot enzyme-linked immunosorbent assay of alkali-solubilized proteins that were not treated with proteases suggested the toxic 25a and 25b proteins were proteolytically derived from a larger molecule of about 28,000 daltons. Alkali-solubilized proteins from an acrystalliferous strain of B. thuringiensis subsp. israelensis and from B. thuringiensis subsp. kurstaki failed to cross-react with antibodies to the 25a protein.     

The inactivation of Bacillus thuringiensis subsp. israelensis toxin by mosquito larvae proteases liberated into the medium

Crystal serine-proteases of B. thuringiensis subsp. israelensis were able to process the 28,000-dalton protein during crystal solubilization. On the other hand, solubilized crystal proteins were degraded during the larvicidal bioassay by the action of serine-proteases liberated by mosquito larvae into the medium, with loss of toxicity. However, proteins in intact crystals were protected from the action of these proteases. This resistance to degradation of crystals partly explains the observation that they are more toxic than solubilized crystal proteins.     

Minireplicon from pBtoxis of Bacillus thuringiensis subsp. israelensis

A 2.2-kb fragment containing a replicon from pBtoxis, the large plasmid that encodes the insecticidal endotoxins of Bacillus thuringiensis subsp. israelensis, was identified, cloned, and sequenced. This fragment contains cis elements, including iterons, found in replication origins of other large plasmids and suggests that pBtoxis replicates by a type A theta mechanism. Two genes, pBt156 and pBt157, encoding proteins of 54.4 kDa and 11.8 kDa, respectively, were present in an operon within this minireplicon, and each was shown by deletion analysis to be essential for replication. The deduced amino acid sequences of the 54.4-kDa and 11.8-kDa proteins showed no substantial homology with known replication (Rep) proteins. However, the 54.4-kDa protein contained a conserved FtsZ domain, and the 11.8 kDa protein contained a helix-turn-helix motif. As FtsZ proteins have known functions in bacterial cell division and the helix-turn-helix motif is present in Rep proteins, it is likely that these proteins function in plasmid replication and partitioning. The minireplicon had a copy number of two or three per chromosome equivalent in B. thuringiensis subsp. israelensis but did not replicate in B. cereus, B. megaterium, or B. subtilis. A plasmid constructed to synthesize large quantities of the Cry11A and Cyt1A endotoxins demonstrated that this minireplicon can be used to engineer vectors for cry and cyt gene expression.     

Minireplicon from pBtoxis of Bacillus thuringiensis subsp. israelensis

A 2.2-kb fragment containing a replicon from pBtoxis, the large plasmid that encodes the insecticidal endotoxins of Bacillus thuringiensis subsp. israelensis, was identified, cloned, and sequenced. This fragment contains cis elements, including iterons, found in replication origins of other large plasmids and suggests that pBtoxis replicates by a type A theta mechanism. Two genes, pBt156 and pBt157, encoding proteins of 54.4 kDa and 11.8 kDa, respectively, were present in an operon within this minireplicon, and each was shown by deletion analysis to be essential for replication. The deduced amino acid sequences of the 54.4-kDa and 11.8-kDa proteins showed no substantial homology with known replication (Rep) proteins. However, the 54.4-kDa protein contained a conserved FtsZ domain, and the 11.8 kDa protein contained a helix-turn-helix motif. As FtsZ proteins have known functions in bacterial cell division and the helix-turn-helix motif is present in Rep proteins, it is likely that these proteins function in plasmid replication and partitioning. The minireplicon had a copy number of two or three per chromosome equivalent in B. thuringiensis subsp. israelensis but did not replicate in B. cereus, B. megaterium, or B. subtilis. A plasmid constructed to synthesize large quantities of the Cry11A and Cyt1A endotoxins demonstrated that this minireplicon can be used to engineer vectors for cry and cyt gene expression.     

Genome Stability of Bacillus thuringiensis subsp. israelensis Isolates

Swedish soil isolates biochemically classified as Bacillus thuringiensis subsp. israelensis were further examined for genetic diversity by multilocus enzyme electrophoresis (MLEE), random amplified polymorphic DNA analysis (RAPD), pulse field gel electrophoresis (PFGE), and Southern blotting, and were compared with reference strains. All the tested strains belonging to the Bt. israelensis serotype H14 were found to be identical, as judged from the RAPD analysis. MLEE analysis gave a similar result; only one H14 strain was found to differ from the remaining H14 strains by one null allele. PFGE analysis confirmed a very close relationship between the H14 strains but revealed an SfiI restriction fragment of variable size. Southern blot analyses were carried out with probes for the chromosomally encoded flagellin gene(s) and the plasmid-encoded mosquitocidal toxins. All probes gave similar hybridization patterns in the H14 strains. The mosquito toxin probes hybridized only to the H14 strains, except for one probe hybridizing to strain 6:3, which was originally isolated from the same soil sample as strains 6:11 and 6:12. Because the RAPD, MLEE, and PFGE analyses showed that strain 6:3 appears to be unrelated to strains 6:11 and 6:12, the presence of a mosquito toxin sequence in strain 6:3 may suggest that gene transfer has occurred. Received: 8 July 1999 / Accepted: 9 August 1999     

Stability of the larvicidal activity of Bacillus thuringiensis subsp. israelensis: amino acid modification and denaturants

The Bacillus thuringiensis subsp. israelensis mosquito larvicidal toxin is not a sulfhydryl-activated toxin. The protein disulfide bonds were cleaved and blocked without loss of toxicity. In contrast, modification of the lysine side chains eliminated toxicity. Additionally, the toxin was resistant to high concentrations of salt (8 M NaBr), organic solvents (40% methanol), denaturants (4 M urea), and neutral detergents (10% Triton X-100). However, it was inactivated by both positively and negatively charged detergents and by guanidine hydrochloride.     

Mosquito larvicidal activity of transgenic Anabaena PCC 7120 expressing toxin genes from Bacillus thuringiensis subsp. israelensis

Genes encoding the mosquito larvicidal toxins Cry4Aa, Cry11Aa, Cyt1Aa and the regulatory P20 from Bacillus thuringiensis subsp. israelensis were introduced into the nitrogen-fixing, filamentous cyanobacterium Anabaena PCC 7120 for expression under control of two strong promoters P(psbA) and P(A1). The clone pRVE4-ADRC displayed toxicity against fourth-instar larvae of Aedes aegypti, the highest ever achieved in cyanobacteria. It was about 2.5-fold more toxic than the respective clone without cyt1Aa [Wu et al., Appl. Environ. Microbiol. 63 (1997) 4971-4975]. Cyt1Aa synergized the combination of Crys by about five-fold. Consistently, the lethal times exerted by pRVE4-ADRC were also reduced (it killed exposed larvae more quickly). This clone may become a useful biological control agent which reduces the probability of resistance development in the target organisms [Wirth et al., Proc. Natl. Acad. Sci. USA 94 (1997) 10536-10540].     

Enduring toxicity of transgenic Anabaena PCC 7120 expressing mosquito larvicidal genes from Bacillus thuringiensis ssp. israelensis

Persistence of biological control agents against mosquito larvae was tested under simulated field conditions. Mosquito larvicidal activity of transgenic Anabaena PCC 7120 expressing cry4Aa, cry11Aa and p20 from Bacillus thuringiensis ssp. israelensis was greater than B. thuringiensis ssp. israelensis primary powder (fun 89C06D) or wettable powder (WP) (Bactimos products) when either mixed with silt or exposed to sunlight outdoors. Reduction of Bactimos primary powder toxicity was at least 10-fold higher than Anabaena's after mixing with silt. In outdoors experiments, Bactimos WP remained toxic (over 30% mortality of 3rd instar Aedes aegypti larvae) for 2-4 days only, while transgenic Anabaena's toxicity endured 8-21 days.     

Stability of the larvicidal activity of Bacillus thuringiensis subsp. israelensis: amino acid modification and denaturants.

The Bacillus thuringiensis subsp. israelensis mosquito larvicidal toxin is not a sulfhydryl-activated toxin. The protein disulfide bonds were cleaved and blocked without loss of toxicity. In contrast, modification of the lysine side chains eliminated toxicity. Additionally, the toxin was resistant to high concentrations of salt (8 M NaBr), organic solvents (40% methanol), denaturants (4 M urea), and neutral detergents (10% Triton X-100). However, it was inactivated by both positively and negatively charged detergents and by guanidine hydrochloride.     

Effect of removal of the cytolytic factor of Bacillus thuringiensis subsp. israelensis on mosquito toxicity

Solubilized crystal protein of Bacillus thuringiensis subsp. israelensis was fractionated by affinity chromatography using a monoclonal antibody directed against the crystal's 28 kDa peptide. The 28 kDa peptide was found to be relatively nontoxic to mosquito larvae although it does contain the hemolytic activity of the crystals. The crystal protein fraction depleted of the 28 kDa peptide was found to be nonhemolytic and to retain nearly full toxicity to mosquito larvae. These results suggest that the 28 kDa peptide is not required for the toxicity of solubilized crystal protein.     

Cloning and characterization of a cytolytic and mosquito larvicidal delta-endotoxin from Bacillus thuringiensis subsp. darmstadiensis

The cytolytic delta-endotoxin gene from Bacillus thuringiensis subsp. darmstadiensis was amplified from genomic DNA by PCR by using primers designed from the sequence of cyt2Aa1 gene of B. thuringiensis subsp. kyushuensis. DNA sequence analysis revealed an open reading frame translating to a 259-amino acid sequence. The cloned gene was designated cyt2Aa2. This gene was highly expressed in Escherichia coli as inclusion bodies that could be solubilized in 50 m M Na(2)CO(3), pH 10.5. Activation of the solubilized protoxin by proteinase K (1% wt/wt, proteinase K/protoxin) yielded the active fragment of about 23 kDa. Cyt2Aa2 showed high hemolytic activity against sheep erythrocytes (hemolytic end- point 0.25 microgram/ml) and was toxic to Culex quinquefasciatus and Aedes aegypti larvae (LC(50) 0.5-1.0 microgram/ml).     

Decreased toxicity of Bacillus thuringiensis subsp. israelensis to mosquito larvae after contact with leaf litter

Bacillus thuringiensis subsp. israelensis is a bacterium producing crystals containing Cry and Cyt proteins, which are toxic for mosquito larvae. Nothing is known about the interaction between crystal toxins and decaying leaf litter, which is a major component of several mosquito breeding sites and represents an important food source. In the present work, we investigated the behavior of B. thuringiensis subsp. israelensis toxic crystals sprayed on leaf litter. In the presence of leaf litter, a 60% decrease in the amount of Cyt toxin detectable by immunology (enzyme-linked immunosorbent assays [ELISAs]) was observed, while the respective proportions of Cry toxins were not affected. The toxicity of Cry toxins toward Aedes aegypti larvae was not affected by leaf litter, while the synergistic effect of Cyt toxins on all B. thuringiensis subsp. israelensis Cry toxins was decreased by about 20% when mixed with leaf litter. The toxicity of two commercial B. thuringiensis subsp. israelensis strains (VectoBac WG and VectoBac 12AS) and a laboratory-produced B. thuringiensis subsp. israelensis strain decreased by about 70% when mixed with leaf litter. Taken together, these results suggest that Cyt toxins interact with leaf litter, resulting in a decreased toxicity of B. thuringiensis subsp. israelensis in litter-rich environments and thereby dramatically reducing the efficiency of mosquitocidal treatments.     

Production of the bioinsecticide Bacillus thuringiensis subsp. israelensis with deltamethrin increases toxicity towards mosquito larvae

Bacillus thuringiensis subsp. israelensis is a bioinsecticide used for larval mosquito control and it represents a safe alternative to chemical insecticides. Despite its environmental safety, it is less efficient and persistent than chemical insecticides. To bypass these limitations, we propose to combine the advantages of chemical and biological insecticides by producing Bti in a medium supplemented with a chemical insecticide (DDT, deltamethrin, permethrin, propoxur or temephos). Among the investigated insecticides, the addition of deltamethrin in the medium induced a higher toxicity (over 6·72‐fold) of the composite deltamethrin‐Bti towards mosquito larvae as compared to Bti alone. This was mainly due to the insertion of deltamethrin into the membranes of Bti spores, as evidenced by a quantification of membrane‐extracted deltamethrin by HPLC. This composite larvicide is a promising tool to decrease the quantity of chemicals dispersed in the environment, to increase the efficacy of Bti and to facilitate its widespread use as a transition between chemical and biological insecticides. Further experiments are required to characterize the mechanisms that underline the incorporation of deltamethrin into Bti to optimize the production and the toxicity of this composite larvicide.

Significance and Impact of the Study

This study is the first report of an increased efficacy of the mosquitocidal bioinsecticide Bacillus thuringiensis subsp. israelensis (Bti) when produced with a chemical insecticide. The results clearly demonstrate that deltamethrin is able to synergize the insecticidal activity of Bti through inclusion into spore membranes, reducing off‐target and nonspecific toxicity occurring when the chemical is used alone as sprays. This new composite chemical–biological insecticide can become an invaluable tool as an intermediate between single chemical usage and the widespread use of Bti, notably in developing countries with limited financial resources for intensive mosquito control campaigns.