Deletion analysis of the 51-kilodalton protein of the Bacillus sphaericus 2362 binary mosquitocidal toxin: construction of derivatives equivalent to the larva-processed toxin.

Bacillus sphaericus 2362 produces a binary toxin consisting of 51- and 42-kDa proteins, both of which are required for toxicity to mosquito larvae. Upon ingestion by larvae, these proteins are processed to 43 and 39 kDa, respectively. Using site-directed mutagenesis, we have obtained N- and C-terminal deletions of the 51-kDa protein and expressed them in B. subtilis by using the subtilisin promoter. Removal of 21 amino acids from the N terminus and 53 amino acids from the C terminus resulted in a protein with the same electrophoretic properties as the 43-kDa degradation product which accumulates in the guts of mosquito larvae. This protein was toxic only in the presence of the 42-kDa protein. A deletion of 32 amino acids at the N terminus combined with a 53-amino-acid deletion at the C terminus resulted in a protein which retained toxicity. Toxicity was lost upon a further deletion of amino acids at potential chymotrypsin sites (41 at the N terminus, 61 at the C terminus). Comparison of the processing of the 51- and the 42-kDa proteins indicated that in spite of their sequence similarity proteolysis occurred at different sites.     

Bacillus sphaericus as a mosquito pathogen: properties of the organism and its toxins.

In the course of sporulation, Bacillus sphaericus produces an inclusion body which is toxic to a variety of mosquito larvae. In this review we discuss the general biology of this species and concentrate on the genetics and physiology of toxin production and its processing in the midgut of the larval host. The larvicide of B. sphaericus is unique in that it consists of two proteins of 51 and 42 kDa, both of which are required for toxicity to mosquito larvae. There is a low level of sequence similarity between these two proteins, which differ in their sequences from all the other known insecticidal proteins of Bacillus thuringiensis. Within the midgut the 51- and 42-kDa proteins are processed to proteins of 43 and 39 kDa, respectively. The conversion of the 42-kDa protein to a 39-kDa protein results in a major increase in toxicity; the significance of the processing of the 51-kDa protein is not known. In contrast to the results with mosquito larvae, the 39-kDa protein is alone toxic for mosquito-derived tissue culture-grown cells, and this toxicity is not affected by the 51-kDa protein or its derivative, the 43-kDa protein. Comparisons of larvae from species which differ in their susceptibility to the B. sphaericus toxin indicate that the probable difference resides in the nature of the target sites of the epithelial midgut cells and not in uptake or processing of the toxin. A similar conclusion is derived from experiments involving tissue culture-grown cells from mosquito species which differ in their susceptibility to the B. sphaericus toxin.     

Modification of the Bacillus sphaericus 51- and 42-kilodalton mosquitocidal proteins: effects of internal deletions, duplications, and formation of hybrid proteins

The 51- and 42-kDa proteins which constitute the binary mosquitocidal toxin of Bacillus sphaericus 2362 have a low overall sequence similarity but share several regions of near identity (L. Baumann, A. H. Broadwell, and P. Baumann, J. Bacteriol. 170:2045-2050, 1988). By using site-directed mutagenesis, deletions of 6 to 16 amino acids in three of these regions of the 51- and 42-kDa proteins were made, and the modified proteins were expressed in Bacillus subtilis. Deletions in both of these proteins resulted in a loss of toxicity for mosquito larvae. Hybrid proteins containing exchanged fragments of the 51- and 42-kDa proteins were inactive when tested in a variety of combinations, thereby indicating that potentially analogous fragments of these two proteins were not functionally equivalent. An internal duplication of 73 amino acids in the 51-kDa protein and 72 amino acids in the 42-kDa protein resulted in a major reduction in toxicity. These results indicate that the conserved regions of the 51- and 42-kDa proteins are necessary for toxicity to larvae and that the 51- and 42-kDa proteins, despite their sequence similarity, are unique, differing from each other by at least one essential attribute.     

Modification of the Bacillus sphaericus 51- and 42-kilodalton mosquitocidal proteins: effects of internal deletions, duplications, and formation of hybrid proteins.

The 51- and 42-kDa proteins which constitute the binary mosquitocidal toxin of Bacillus sphaericus 2362 have a low overall sequence similarity but share several regions of near identity (L. Baumann, A. H. Broadwell, and P. Baumann, J. Bacteriol. 170:2045-2050, 1988). By using site-directed mutagenesis, deletions of 6 to 16 amino acids in three of these regions of the 51- and 42-kDa proteins were made, and the modified proteins were expressed in Bacillus subtilis. Deletions in both of these proteins resulted in a loss of toxicity for mosquito larvae. Hybrid proteins containing exchanged fragments of the 51- and 42-kDa proteins were inactive when tested in a variety of combinations, thereby indicating that potentially analogous fragments of these two proteins were not functionally equivalent. An internal duplication of 73 amino acids in the 51-kDa protein and 72 amino acids in the 42-kDa protein resulted in a major reduction in toxicity. These results indicate that the conserved regions of the 51- and 42-kDa proteins are necessary for toxicity to larvae and that the 51- and 42-kDa proteins, despite their sequence similarity, are unique, differing from each other by at least one essential attribute.     

Cytotoxicity and ADP-ribosylating activity of the mosquitocidal toxin from Bacillus sphaericus SSII-1: possible roles of the 27- and 70-kilodalton peptides.

Clones expressing regions of the 100-kDa Bacillus sphaericus SSII-1 mosquitocidal toxin (Mtx) as fusion proteins with glutathione S-transferase were constructed, and the toxin-derived peptides were purified. The in vitro ADP-ribosylation activities of these peptides and their effects on larvae and cells in culture were studied. Mtx25 (amino acids 30 to 493) was found to ADP-ribosylate two proteins with molecular masses of 38 and 42 kDa, respectively, in Culex quinquefasciatus (G7) cell extracts, in addition to ADP-ribosylating itself. Mtx21 (amino acids 30 to 870; or a combination of Mtx25 and Mtx26 (amino acids 259 to 870) caused mortality in C. quinquefasciatus larvae. Mtx25, Mtx26, or Mtx24 (amino acids 30 to 276) alone and Mtx24 in combination with Mtx26 were not toxic to larvae. Mtx21 and Mtx26 produced marked morphological changes in G7 cells and to a lesser extent in Aedes aegypti cells but had no effect on Anopheles gambiae or HeLa cells. Thus, a domain in the N-terminal region of the Mtx protein is sufficient for ADP-ribosylation of C. quinquefasciatus cell protein, and a domain in the C-terminal region is sufficient for toxicity to cultured C. quinquefasciatus cells; however, both regions are necessary for toxicity to mosquito larvae.     

Cloning of the gene for the larvicidal toxin of Bacillus sphaericus 2362: evidence for a family of related sequences.

During sporulation, Bacillus sphaericus 2362 produces a parasporal crystalline protein which is toxic for the larvae of a number of mosquito species. Using the Escherichia coli cloning vector lambda gt11, in which gene products of the inserts may be fused to beta-galactosidase, we isolated 29 bacteriophages which produced peptides-reacting with antiserum to crystal protein. On the basis of restriction enzyme analyses of the recombinants and Ouchterlony immunodiffusion experiments with induced lysogens as a source of antigens, the recombinants were assigned to three groups, designated A, B, and C. Group A consisted of three clones which appeared to express all or part of the B. sphaericus toxin gene from their own promoters and one clone producing a beta-galactosidase-toxin fusion protein. The host cells of two induced recombinant lysogens of this group were toxic to larvae of Culex pipiens. A cell suspension containing 174 ng (dry weight) of the more toxic recombinant per ml killed 50% of the larvae. Both recombinants formed peptides with molecular sizes of 27, 43, and 63 kilodaltons (kDa). The antigenically related 27- and 43-kDa peptides were distinct from the 63-kDa peptide, which resembled crystals from sporulating cells of B. sphaericus in which antigenically distinct 43- and 63-kDa proteins are derived from a 125-kDa precursor. A 3.5-kilobase HindIII fragment from recombinants having toxic activity against larvae was subcloned into pGEM-3-blue. E. coli cells harboring this fragment were toxic to mosquito larvae and produced peptides of 27, 43, and 63 kDa. The distribution of the A gene among strains of B. sphaericus of different toxicities suggested that it is the sole or principal gene encoding the larvicidal crystal protein. The two recombinants of group B and the 23 of group C were all beta-galactosidase fusion proteins, suggesting that in E. coli these genes were not readily expressed from their own promoters. The distribution of these two genes in different strains of B. sphaericus suggested that they do not have a role in the toxicity of this species to mosquito larvae.     

Delineation of the minimal portion of the Bacillus sphaericus 1593M toxin required for the expression of larvicidal activity

The two genes of Bacillus sphaericus 1953M coding for the 51.4-kDa and 41.9-kDa proteins are both required for the expression of the active larvicidal toxin in Escherichia coli. The minimal size of the active peptide of the 41.9-kDa toxin was defined by in vitro deletion analysis of the gene and found to consist of 338 amino acids (38.3 kDa). N-terminal deletions past the Ile18 residue and C-terminal deletions past the His352 residue result in the loss of toxic activity and rapid degradation of such modified toxins by host proteases. The minimal active 38.3-kDa peptide produced in E. coli seems to mimick the stable processed form of the toxin found in larval midguts. However, it still requires the action of the synergistic 51.4-kDa protein for the larvicidal activity.     

Genetic determinants of host ranges of Bacillus sphaericus mosquito larvicidal toxins.

The 51.4-kDa-41.9-kDa binary toxin produced by different strains of Bacillus sphaericus shows differential activity toward Culex quinquefasciatus, Aedes atropalpus, and Aedes aegypti mosquito larvae. The patterns of larvicidal activity toward all three mosquito species and growth retardation in A. aegypti have been shown to be due to the 41.9-kDa protein. By using mutant toxins expressed in Escherichia coli, insecticidal activity and growth retardation correlated with amino acids centered around position 100 of the 41.9-kDa protein. In its response to these toxins, A. atropalpus resembled C. quinquefasciatus rather than its congener, A. aegypti.     

Binding of Bacillus sphaericus binary toxin to a specific receptor on midgut brush-border membranes from mosquito larvae.

The presence of specific receptors for Bacillus sphaericus binary toxin on brush-border membrane fractions (BBMF) from Culex pipiens larvae midgut cells was demonstrated by an in vitro binding assay. Both activated and radiolabelled polypeptides from the 51-kDa and 42-kDa binary toxin of B. sphaericus 1593 specifically bound to BBMF. Direct binding and homologous competition experiments indicated a single class of B. sphaericus toxin receptors, with a dissociation constant (Kd) of approximately 20 nM and a maximum binding capacity (Bmax) of approximately 7 pmol/mg BBMF protein. The sugars GalNAc, GlcNAc and N-acetyl neuraminic acid had no detectable inhibitory effect on toxin binding to C. pipiens BBMF. Binding experiments with the non-susceptible mosquito species Aedes aegypti failed to detect significant binding of B. sphaericus binary toxin to A. aegypti BBMF.     

Purification of the larvicidal toxin of Bacillus sphaericus and evidence for high-molecular-weight precursors

Crystals were purified from spore-crystal complexes of Bacillus sphaericus 2362 by disruption in a French pressure cell followed by centrifugation through 48% (wt/vol) NaBr. Crystals from such preparations had a 50% lethal concentration of 6 ng of protein per ml for the larvae of the mosquito Culex pipiens. When subjected to polyacrylamide gel electrophoresis under denaturing conditions, the proteins in B. sphaericus crystals migrated in positions corresponding to 43, 63, 98, 110, and 125 kilodaltons (kDa); solubilization of the crystal at pH 12 with NaOH eliminated all but the bands at 43 and 63 kDa. Since NaOH-solubilized preparations were toxic to mosquito larvae, these proteins were purified to electrophoretic homogeneity and antiserum was obtained to each. Analysis of the two purified proteins indicated that the 43-kDa protein was toxic to mosquito larvae (50% lethal concentration, 35 ng of protein per ml), whereas the 63-kDa protein was not. Further differences between them were their amino acid compositions, their lack of immunological cross-reactivity, their opposite net charges at pH 7.5, and their susceptibility to digestion by larval midgut proteases (the 63-kDa protein was highly susceptible, whereas the 43-kDa protein was not). The sequence of the 40 N-terminal residues of the 43-kDa protein was determined and found to contain a high percentage of hydrophobic amino acids. The sequence of the 63-kDa protein could not be determined, since it had multiple N termini. By electrophoretically separating the crystal proteins and then electroblotting onto nitrocellulose paper and visualizing the bands with antisera to the 43- and 63-kDa proteins in conjunction with an immunoblot assay, it was found that the high-molecular-mass crystal proteins (98 to 125 kDa) contained antigenic determinants of both proteins. These results suggested that the lower-molecular-weight crystal proteins detected in polyacrylamide gels after electrophoresis under denaturing conditions were derivatives of one or more of the higher-molecular-weight crystal proteins. In vivo studies of the products of crystal degradation by larvae of Culex pipiens indicated that the high-molecular-weight proteins and the 63-kDa antigenic determinants were rapidly degraded and that a 40-kDa protein related to the 43-kDa toxin persisted for the duration of the experiment (4 h). Some of the studies performed with B.sphaericus 2362 were extended to strains 1593, 1691, and 2297 of this species with results which indicated a high degree of similarity between the crystal proteins of all these larvicidal strains.     

Effects of components of theBacillus sphaericus toxin on mosquito larvae and mosquito-derived tissue culture-grown cells

Bacillus sphaericus 2362 produces a parasporal crystal containing 42 and 51 kilodalton (kDa) proteins. Both of these proteins are required for toxicity to mosquito larvae; neither is toxic alone. When overexpressed inB. subtilis, these two proteins accumulate as amorphous inclusions (AIs). Bioassays involving larvae ofCulex pipiens and different ratios of these AIs indicated that maximal toxicity was observed at a ratio of approximately one 42-kDa protein to one 51-kDa protein. Purified preparations of these proteins, as well as derivatives similar to those which accumulate in the gut of mosquito larvae, were also toxic when combined, but not toxic singly. Different results were obtained when the toxicity of these preparations was tested for tissue culture-grown cells ofC. quinquefasciatus. Under these conditions, the 39-kDa derivative of the 42-kDa protein was alone sufficient for toxicity, which was not increased by the addition of the 51-kDa protein or its derivatives. These results indicate that theB. sphaericus larvicide acts as a binary toxin in mosquitos, whereas only the 39-kDa protein is required for full toxicity to tissue culture-grown cells.     

Identification of the functional site in the mosquito larvicidal binary toxin of Bacillus sphaericus 1593M by site-directed mutagenesis

To study the mode of action of the binary toxin (51- and 42-kDa) of Bacillus sphaericus, amino acid residues were substituted at selected sites of the N- and C-terminal regions of both peptides. Bioassay results of the mutant binary toxins tested against mosquito larvae, Culex quinquefasciatus, revealed that most of the substitutions made on both peptides led to either decrease or total loss of the activity. Furthermore, receptor binding studies carried out for some of the mutants of the 42-kDa peptide showed mutations in N- and C-terminal regions of the 42-kDa peptide did not affect the binding of the binary toxin to brush border membrane vesicles of mosquito larvae. One of the mutants having a single amino acid substitution at the C-terminal region ((312)R) of the 42-kDa peptide completely abolished the biological activity, implicating the role of this residue in membrane pore formation. These results indicate the importance of the C-terminal region of the 42-kDa of binary toxin, in general, and particularly the residue (312)R for biological activity against mosquito larvae.     

Binding of purified Bacillus sphaericus binary toxin and its deletion derivatives to Culex quinquefasciatus gut: elucidation of functional binding domains.

Highly larvicidal strains of Bacillus sphaericus produce a binary toxin composed of 51 and 42 kDa proteins which binds to sharply delineated regions of the gastric caecum and posterior midgut of susceptible larvae of the mosquito Culex quinquefasciatus. To investigate the role of the individual subunits and the organization of functional binding regions within the toxin, plasmids were constructed for the expression in Escherichia coli of the toxin proteins and their NH2- and COOH-terminal deletion derivatives as fusions with glutathione S-transferase (GST). Toxin proteins were purified by affinity chromatography followed by cleavage from the GST carrier with thrombin. The LC50 values for the purified toxin proteins and their deletion derivatives were determined. The binding patterns of fluorescently labelled toxin suggested that the 51 kDa protein is the primary binding component of the toxin and mediates the regional binding and internalization of the 42 kDa protein. Examination of the toxin deletion derivatives revealed that the NH2-terminal region of the 51 kDa protein was required for binding to the larval gut, whilst the COOH-terminal region was responsible for interacting with the 42 kDa protein. Toxicity was strongly correlated with the subsequent internalization of the toxin, probably by endocytosis.     

Proteolysis in the gut of mosquito larvae results in further activation of the Bacillus sphaericus toxin.

Gut proteases from the larvae of the mosquito Culex pipiens convert the 43-kilodalton (kDa) toxin from Bacillus sphaericus 2362 to a 40-kDa peptide. The 50% lethal concentration of this peptide for tissue culture-grown cells of Culex quinquefasciatus was 1.0 microgram/ml (as determined by the intracellular ATP assay), 54-fold less than that of the 43-kDa peptide. Gut proteases from Anopheles gambiae and Aedes aegypti, as well as bovine pancreatic trypsin, also converted the 43-kDa protein to a 40-kDa peptide which was indistinguishable from the peptide formed by the proteases from C. pipiens with respect to its toxicity to tissue culture-grown cells of C. quinquefasciatus. Evidence for the in vivo conversion of the 43-kDa protein to the 40-kDa peptide was also obtained from experiments in which larvae of C. pipiens, Anopheles gambiae, and Aedes aegypti were fed crystals from B. sphaericus 2362. By using the exclusion of trypan blue as an indication of cell viability, it was shown that chitobiose, chitotriose, N-acetylmuramic acid, and N-acetylneuraminic acid decreased the toxicity of the 40-kDa peptide (from 100 to 50% mortality at about 10 mM concentrations of these sugars). Muramic acid, N-acetylgalactosamine, and N-acetylglucosamine were less effective, while several sugars had no effect, suggesting that the 40-kDa toxin binds to specific receptors on the cell membrane. The 40-kDa protein was less toxic to tissue culture-grown cells of Anopheles gambiae and Aedes dorsalis, and the same sugars which reduced the toxicity for cells of C. quinquefasciatus were also effective in reduction of toxicity for these cell lines.(ABSTRACT TRUNCATED AT 250 WORDS)     

Proteolysis in the gut of mosquito larvae results in further activation of the Bacillus sphaericus toxin

Gut proteases from the larvae of the mosquito Culex pipiens convert the 43-kilodalton (kDa) toxin from Bacillus sphaericus 2362 to a 40-kDa peptide. The 50% lethal concentration of this peptide for tissue culture-grown cells of Culex quinquefasciatus was 1.0 microgram/ml (as determined by the intracellular ATP assay), 54-fold less than that of the 43-kDa peptide. Gut proteases from Anopheles gambiae and Aedes aegypti, as well as bovine pancreatic trypsin, also converted the 43-kDa protein to a 40-kDa peptide which was indistinguishable from the peptide formed by the proteases from C. pipiens with respect to its toxicity to tissue culture-grown cells of C. quinquefasciatus. Evidence for the in vivo conversion of the 43-kDa protein to the 40-kDa peptide was also obtained from experiments in which larvae of C. pipiens, Anopheles gambiae, and Aedes aegypti were fed crystals from B. sphaericus 2362. By using the exclusion of trypan blue as an indication of cell viability, it was shown that chitobiose, chitotriose, N-acetylmuramic acid, and N-acetylneuraminic acid decreased the toxicity of the 40-kDa peptide (from 100 to 50% mortality at about 10 mM concentrations of these sugars). Muramic acid, N-acetylgalactosamine, and N-acetylglucosamine were less effective, while several sugars had no effect, suggesting that the 40-kDa toxin binds to specific receptors on the cell membrane. The 40-kDa protein was less toxic to tissue culture-grown cells of Anopheles gambiae and Aedes dorsalis, and the same sugars which reduced the toxicity for cells of C. quinquefasciatus were also effective in reduction of toxicity for these cell lines.(ABSTRACT TRUNCATED AT 250 WORDS)     

The receptor of Bacillus sphaericus binary toxin in Culex pipiens (Diptera: Culicidae) midgut: molecular cloning and expression

Culex pipiens larval midgut is the primary target of the binary toxin (Bin) present in parasporal inclusions of Bacillus sphaericus. Cpm1, a 60-kDa protein purified from brush border membranes, has been proposed as the receptor of the Bin toxin in the midgut epithelial cells of mosquitoes. We have cloned and characterized the corresponding cDNA from midgut of Culex pipiens larvae. The open reading frame predicted a 580 amino-acid protein with a putative signal peptide at the N-terminus and a putative GPI-anchoring signal at the C-terminus. The amino acid sequence of the cloned Cpm1 exhibited 39-43% identities with insect maltases (alpha-glucosidases and alpha-amylases). Recombinant Cpm1 expressed in E. coli specifically bound to the Bin toxin and had a significant alpha-glucosidase activity but no alpha-amylase activity. These results support the view that Cpm1 is an alpha-glucosidase expressed in Culex midgut where it constitutes the receptor for the Bin toxin. To date, this is the first component involved in the mosquitocidal activity of the Bacillus sphaericus Bin toxin to be characterized. Its identification provides a key step to elucidate the mode of action of the Bin toxin and the mechanisms of resistance developed against it by some mosquito strains.     

An analysis of the genes encoding the 51.4- and 41.9-kDa toxins of Bacillus sphaericus 2297 by deletion mutagenesis: the construction of fusion proteins

A series of deletion mutants have been constructed, in which varying numbers of amino acids have been deleted from both the N- and C-termini of both the 51.4- and 41.9-kDa toxins of Bacillus sphaericus. The results show that between 34-39 and 52-54 amino acids respectively at the N- and C-termini of the 51.4-kDa protein, are not essential for toxicity. In the case of the 41.9-kDa protein, the removal of only 7 amino acids from the C-terminus abolishes toxicity whilst at least 17 amino acids can be deleted from the N-terminus without loss of toxicity. A fusion protein with the 51.4-kDa derived sequence N-terminal to the 41.9-kDa sequence yielded a protein of Mr 87 kDa which was not toxic by itself. When supplemented with cells expressing only the 51.4-kDa protein, toxicity was restored. In contrast, another fusion protein, in which the gene order was reversed, was shown to be fully active in toxicity assays.     

Role of the gut proteinases from mosquito larvae in the mechanism of action and the specificity of the Bacillus sphaericus toxin

Gut proteinases from larvae of mosquito species both susceptible and not susceptible to Bacillus sphaericus converted the 43-kDa toxin to a 40-kDa polypeptide exhibiting enhanced cytotoxicity to mosquito cell cultures. The toxin was also activated by gut proteinases from the nonsusceptible Lepidoptera Spodoptera littoralis in vitro and in vivo. Therefore, the specificity of Bacillus sphaericus toxin does not seem to be determined by gut proteinase action. However, susceptibility of mosquito cell cultures did not reflect the specificity of the toxin, which must now be investigated at the cellular level in the larvae.     

Activation pattern and toxicity of the Cry11Bb1 toxin of Bacillus thuringiensis subsp. medellin

Bacillus thuringiensis protoxins undergo proteolytic processing in the midgut of susceptible insects to become active. The ability to process the Cry11Bb1 protoxin by trypsin and Culex quinquefasciatus larval gut extracts was tested. The protease activity indicated by the appearance of proteolytic products increased with an increment in pH, with the highest activity being observed at pH 10.6. A time course study showed the proteolysis of the 94-kDa Cry11Bb protein ending with the production of fragments of relative molecular mass of 30 and 35 kDa within 5 min. In vitro, gut proteases extract cleaved the solubilized toxin between Ser59 and Ile60 and between Ala395 and Asn396, generating a 30-kDa N-terminal and a 35-kDa C-terminal fragment, respectively. Similarly, mosquito larvae processed in vivo the parasporal inclusions, generating the same fragments as those observed in vitro. The Cry11Bb1 protoxin activated with trypsin or gut proteases showed larvicidal activity against C. quinquefasciatus first instar larvae. The data suggest that gut proteases participate in the activation of CryllBbl protoxin, generating at least two different fragments on which the activity could reside.     

The 42- and 51-kilodalton mosquitocidal proteins of Bacillus sphaericus 2362: construction of recombinants with enhanced expression and in vivo studies of processing and toxicity.

After site-directed mutagenesis, the genes coding for the 42- and 51-kilodalton (kDa) mosquitocidal proteins of Bacillus sphaericus 2362 were placed under the regulation of the aprE (subtilisin) promoter of the Bacillus subtilis vector pUE (a derivative of pUB18). The levels of expression of the gene products in B. subtilis DB104 and B. sphaericus 718 were assessed by bioassays with larvae of Culex pipiens and by Western immunoblots. The results indicated that a higher amount of protein was produced in B. subtilis DB104. Electron microscopic examination of B. subtilis DB104 and B. sphaericus 718 containing the 42- and 51-kDa proteins indicated that amorphous inclusions accumulated in the former species and that crystals identical in appearance to that found in B. sphaericus 2362 were produced in the latter. Strains producing only the 42- or the 51-kDa protein were not toxic to larvae of C. pipiens. A mixture of both strains, a single strain producing both proteins, or a fusion of the 51- and the 42-kDa proteins was toxic. The amount of B. subtilis DB104 containing the 42- and the 51-kDa proteins necessary to kill 50% of the larvae of C. pipiens was 5.6 ng (dry weight) of cells per ml. This value was significantly lower than that for B. sphaericus 2362 (14 ng [dry weight] per ml). Larvae consuming purified amorphous inclusions containing the 42-kDa protein degraded this protein this protein to primarily 39- and 24-kDa peptides, whereas inclusions with the 51-kDa protein were primarily degraded to a protein of 44 kDa. Past studies involving purified proteins from B. sphaericus 2362 indicate an associate of toxicity with the 39-kDa peptide. The results presented here suggest that the 44-kDa degradation product of the 51-kDa protein may also be required for toxicity.