Demonstration of a cholesterol-dependent cytolysin in a noninsecticidal Bacillus sphaericus strain and evidence for widespread distribution of the toxin within the species

During the course of screening Bacillus species from food and water in Norway, we isolated a strain of Bacillus sphaericus of DNA homology group V, not previously recognized to contain entomopathogenic strains, that was cytotoxic to Vero cell epithelia. Peptide mass fingerprinting of a protein purified from the culture supernatant of B. sphaericus B354 identified a cholesterol-dependent cytolysin (CDC) with high amino acid sequence identity with sphaericolysin, a CDC identified recently in B. sphaericus DNA homology group IIA. The toxin was haemolytic against erythrocytes from a range of species. Haemolysis was potentiated by dithiothreitol and inhibited by preincubation with cholesterol. The toxin induced lactate dehydrogenase release from Vero cells and formed pores in planar lipid bilayers. The distribution of CDC genes in B. sphaericus was examined, with CDC gene products obtained in 13 out of 17 strains representing four of the six DNA homology groups. Thus, we demonstrate the presence of a CDC in a nonentomopathogenic DNA homology group of B. sphaericus (group V) with typical CDC characteristics. CDCs appear to be present in a high proportion of B. sphaericus strains and are not restricted to group IIA insecticidal strains.     

Evaluation of an rRNA-targeted oligonucleotide probe for the detection of mosquitocidal strains of Bacillus sphaericus in soils: characterization of novel strains lacking toxin genes

Abstract: Colonies of Bacillus sphaericus on primary isolation have been identified using an oligonucleotide probe targeted to a specific region of the 16S rRNA. Of 3440 colonies from soil samples from Brazil, 57 hybridized to the probe in colony blots but when purified DNA was used in slot-blots the probe was more specific and only 27 isolates hybridized. Of these, 20 strains were confirmed as members of DNA homology group IIA (potential mosquitocidal strains) by ribotyping and isoenzyme analysis. However, none of these strains was toxic to Anopheles or Culex larvae, nor did they contain recognized toxin genes. This is the first demonstration of such non-pathogenic strains of B. sphaericus DNA homology group IIA and their common occurrence suggests that pathogenicity is not an important contribution to the success of these bacteria in the environment. Similar screening of strains from Scottish soils indicated that B. sphaericus DNA homology group IIA strains were less common in this habitat and none were recovered on this occasion.     

DNA fingerprinting of the mosquito pathogen Bacillus sphaericus with M13 DNA as a probe

Hypervariable nucleotide sequences were detected in Bacillus sphaericus by hybridization with radioactively labelled M13 DNA. Different serotypes could be distinguished by their hybridization profiles. The appearance of bands common for mosquito-pathogenic strains and their absence in an apathogenic strain opens the probability that M13 could hybridize to specific alleles, related to insect toxicity.     

Comparison of the cellular fatty acid composition of a bacterium isolated from a human and alleged to be Bacillus sphaericus with that of Bacillus sphaericus isolated from a mosquito larvicide.

The cellular fatty acid (CFA) composition of the cytoplasmic membrane of a bacillus isolated from a human lung and deposited in the National Collection of Type Cultures as Bacillus sphaericus NCTC 11025 was determined by gas-liquid chromatography. The CFA composition of B. sphaericus 2362, isolated from a microbial larvicide, and those of B. sphaericus reference strains obtained from public collections were also determined. Samples were grouped by hierarchical cluster analysis based on the unpaired-group method using arithmetic averages. Samples that linked at a Euclidean distance of < or = 2.0 U were considered to belong to the same strain. NCTC 11025 and the type strain of B. sphaericus, ATCC 14577, were mixed; all other isolates were monotypic. The predominant fatty acid in NCTC 11025 was 12-methyltetradecanoic acid, while the predominant fatty acid in the remaining isolates was 13-methyltetradecanoic acid. NCTC 11025 linked to the other isolates at a Euclidean distance of 83.8 U, and we concluded that it belongs to a different species that we could not identify. We could distinguish among six DNA homology groups of B. sphaericus by using fatty acids. Within DNA homology group IIA, strain 2362 could be distinguished from other strains belonging to serotype H5a, 5b. We concluded that CFA analysis is a useful technique to determine if future human isolates identified as B. sphaericus in fact belong to other species of bacteria or whether the isolates originated from commercial products.     

Phylogenetic analysis and heterologous expression of surface layer protein SlpC of Bacillus sphaericus C3-41

The surface layer protein encoding genes from five mosquito-pathogenic Bacillus sphaericus isolates were amplified and sequenced. Negative staining of the S-layer protein extracted from the cell wall of wild-type B. sphaericus C3-41 was prepared. It showed a flat-sheet crystal lattice structure. Two genes encoding the entire and N-terminally truncated S-layer protein (slpC and DeltaslpC respectively), were ligated into plasmid pET28a and expressed in Escherichia coli. SDS-PAGE revealed that about 130 KD and 110 KD proteins could be expressed in the cytoplasm of recombinant E. coli BL21(pET28a/slpC) and E. coli BL21(pET28a/DeltaslpC) respectively. Furthermore, an intracellular sheet-like or fingerprint-shape structure was investigated in two recombinant strains, which expressed SlpC and DeltaSlpC protein respectively, by ultrathin microscopy study, but bioassay results suggested that the S-layer protein of wild B. sphaericus C3-41 and recombinant E. coli BL21 (pET28a/slpC) have no direct toxicity against mosquito larvae. These results should provide information for further understanding of the function of S-layer protein of pathogenic B. sphaericus.     

Efficient expression of a 100-kilodalton mosquitocidal toxin in protease-deficient recombinant Bacillus sphaericus.

The expression of the 100-kDa mosquitocidal toxin (Mtx) during vegetative growth and sporulation in nine different mosquito-larvicidal strains of Bacillus sphaericus has been analyzed. In five out of the nine strains the 100-kDa toxin was found to be expressed predominantly in the vegetative phase of growth, and in all nine strains the level of the toxin in sporulated cells was very low or undetectable. Strains in four out of the six DNA homology groups of B. sphaericus produced intracellular and extracellular proteases, which degraded the 100-kDa toxin, during sporulation. The 100-kDa toxin gene was expressed by using its native promoter on a multicopy number plasmid in B. sphaericus 1693 (protease negative) and B. sphaericus 13052 (protease positive). High levels of the 100-kDa toxin were produced in vegetative cells of both strains as well as in sporulated cells of protease-negative strain 1693, which is in contrast to the low levels of the 100-kDa toxin produced in sporulated cells of protease-positive strain 13052. Thus, the small amount of the 100-kDa toxin in sporulated cells of the nine mosquito-larvicidal strains is probably due to degradation of the 100-kDa toxin synthesized during vegetative growth by a protease(s) produced during sporulation. B. sphaericus 1693 transformed with the 100-kDa toxin gene was as toxic to mosquito larvae during both vegetative growth and sporulation as the natural high-toxicity strains of sporulated B. sphaericus.(ABSTRACT TRUNCATED AT 250 WORDS)     

Isolation and identification of Bacillus sphaericus strains pathogenic for mosquito larvae

Three selective media for the isolation of Bacillus sphaericus have been compared. BATS medium and a formulation employing adenosine as the principal carbon source were the most effective for the recovery of spores of strain 1593. Anthranilic acid as the principal carbon source was less efficient. Eighty-four strains were isolated from mud samples using these media and were identified by computer. Identifications were confirmed for representative strains using DNA sequence homology. Most were B. sphaericus sensu stricto or members of an unnamed group. However, one strain (BSE 18) was identified as the DNA homology group IIB and this organism was found to be highly toxic toward larvae of Culex pipiens. Southern hybridization of BSE 18 DNA to a probe prepared from the cloned toxin gene from strain 1593 revealed that BSE 18 contained a typical gene for the 41.9-kDa toxin.     

Characterization of the surface protein layers of the mosquito-pathogenic strains of Bacillus sphaericus.

The protein surface layers on the cell walls of mosquito-pathogenic and nonpathogenic Bacillus sphaericus strains were studied by structural, biochemical, and serological methods. The surface structure of two representative insect-pathogenic strains had the form of a delicate linear array with a repeat interval of 5 nm. This was distinctly different from the tetragonal array of the P-1 strain in spacing and arrangement. The surface layers were composed of acidic glycoproteins with molecular weights ranging from approximately 133,000 to 155,000. Peptide mapping and serological analysis of the surface proteins revealed eight distinct groups among the pathogens. These groups were very similar to the groupings determined by flagellar-antigen serotyping and bacteriophage typing.     

New gene from nine Bacillus sphaericus strains encoding highly conserved 35.8-kilodalton mosquitocidal toxins.

A new gene encoding a 35.8-kDa mosquitocidal toxin (Mtx3; 326 amino acids) was isolated from Bacillus sphaericus SSII-1 DNA. Mtx3 is a new type of mosquitocidal toxin with homology to the Mtx2 mosquitocidal toxin of B. sphaericus SSII-1, the epsilon-toxin of Clostridium perfringens, and the cytotoxin of Pseudomonas aeruginosa. The mtx3 gene is highly conserved and widely distributed in both high- and low-toxicity mosquito larvicidal strains of B. sphaericus.     

Novel cost-effective method of screening soils for the presence of mosquito-pathogenic bacilli

AIMS: The aim was to simplify the cumbersome conventional process of isolating virulent bacilli, which involves isolating all bacilli strains from a source followed by screening for strains that are effective for bio-control of mosquito vectors. METHODS: A new simplified technique involving eight steps was devised for screening soil samples for the presence of mosquito-pathogenic bacilli before isolating individual strains. RESULTS: Using the new technique, we obtained eight bacilli strains (KSD1-8) showing pathogenic activity against mosquito larvae from three out of 10 soil samples screened. These strains were characterized, identified and the main bioassay tests were performed with three most promising strains (KSD-4, KSD-7 and KSD-8), and their pathogenic activity against Anopheles stephensi Liston, Culex quinquefasciatus, Say and Aedes aegypti Linnaeus compared well with commercial reference strains of B. thuringiensis israelensis and B. sphaericus. SIGNIFICANCE AND IMPACT OF THE STUDY: The new technique of screening soil samples for the presence of virulent pathogenic strains of bacilli against mosquito larvae proved quick, efficient and cost effective.     

An enzyme linked immunosorbent assay for monitoring Bacillus sphaericus toxin.

Larvicidal proteins of B. sphaericus H5a5b (strain VCRC B42), purified by ion-exchange chromatography were used to raise antibodies in rabbits. The antibodies were specific in reacting to alkali-solubilized fractions from whole cells of toxic strains only. Ouchterlony immunodiffusion showed homology in toxin structure between strains of different serotype. A sandwich ELISA using avidin-biotin amplification was standardized. The lowest detectable limit was 6.25 ng/ml. Near linear quantitative binding of the antigen was found in the range 25-200 ng/ml. The growth, toxin level and LC50 values during various stages of fermentation of B. sphaericus strains 1593 and B42 were compared. There was significant correlation between LC50 values and toxin levels as measured by ELISA.     

Improvement of Bacillus sphaericus toxicity against dipteran larvae by integration, via homologous recombination, of the Cry11A toxin gene from Bacillus thuringiensis subsp. israelensis.

Integrative plasmids were constructed to enable integration of foreign DNA into the chromosome of Bacillus sphaericus 2297 by in vivo recombination. Integration of the aphA3 kanamycin resistance gene by a two-step procedure demonstrated that this strategy was applicable with antibiotic resistance selection. Hybridization experiments evidenced two copies of the operon encoding the binary toxin from B. sphaericus in the recipient strain. The Bacillus thuringiensis subsp. israelensis cry11Aal gene (referred to as cry11A), encoding a delta-endotoxin with toxicity against Culex, Aedes, and Anopheles larvae, was integrated either by a single crossover event [strain 2297 (::pHT5601), harboring the entire recombinant plasmid] or by two successive crossover events [strain 2297 (::cry11A)]. The level of the Cry11A production in B. sphaericus was high; two crystalline inclusions were produced in strain 2297 (::pHT5601). Synthesis of the Cry11A toxin conferred toxicity to the recombinant strains against Aedes aegypti larvae, for which the parental strain was not toxic. Interestingly, the level of larvicidal activity of strain 2297 (::pHT5601) against Anopheles stephensi was as high as that of B. thuringiensis subsp. israelensis and suggested synergy between the B. thuringiensis and B. sphaericus toxins. The toxicities of parental and recombinant B. sphaericus strains against Culex quinquefasciatus were similar, but the recombinant strains killed the larvae more rapidly. The production of the Cry11A toxin in B. sphaericus also partially restored toxicity for C. quinquefasciatus larvae from a population resistant to B. sphaericus 1593. In vivo recombination therefore appears to be a promising approach to the creation of new B. sphaericus strains for vector control.     

Random and repetitive primer amplified polymorphic DNA analysis of Bacillus sphaericus

A comparative study of 15 strains representing the five homology groups of Bacillus sphaericus was performed by two PCR methods: RAPD and rep-PCR fingerprinting. The PCR analysis performed with primers corresponding to the naturally occurring repetitive sequences REP, ERIC and BOX, as well as with three random primers, showed highly variable patterns and allowed differentiation of the strains studied. This demonstrated the high discriminative power of the methods. The cluster analysis revealed a low level of similarity between the different homology groups, and within groups I and III, which is evidence of the high genetic heterogeneity of the species B. sphaericus. Close genetic relatedness was observed for the representatives of group IIA, pathogenic to mosquitoes, which supports the idea for differentiation of this group as a separate species.     

Cloning, sequencing, and expression of a gene encoding a 100-kilodalton mosquitocidal toxin from Bacillus sphaericus SSII-1.

A cosmid library was prepared from a partial BamHI digest of total DNA from Bacillus sphaericus SSII-1. Two hundred fifty Escherichia coli clones were screened for toxicity against larvae of the mosquito Culex quinquefasciatus. One toxic clone, designated pKF2, was chosen for further study. Two toxic subclones, designated pXP33 and pXP34, obtained by ligating PstI-derived fragments of pKF2 into pUC18, contained the same 3.8-kb fragment, but in opposite orientations. Sequence analysis revealed the presence of an open reading frame corresponding to a 100-kDa protein and the 3' end of a further open reading frame having significant homology to open reading frames of transposons Tn501 and Tn21. The sequence of the SSII-1 toxin was compared with those of known toxins and was found to show regional homology to those of ADP-ribosyltransferase toxins. The distribution of the toxin gene among other B. sphaericus strains was examined.     

Distribution of restriction endonucleases among some entomopathogenic strains of Bacillus sphaericus

The restriction enzyme Bsp TI, an isoschizomer of Hae III (recognition site GGCC), has been detected in eight strains of serotype 5a5b and two serotype 3 strains of the entomopathogenic bacterium Bacillus sphaericus . Strains from other serotypes contained the enzymes Bsp TII and Bsp TIII, which digested pBR322 DNA into similar banding patterns after agarose gel electrophoresis but differed in their susceptibility to methylation of the substrate. Strains from serotypes 9, 25 and 26a26b were lacking in restriction enzyme activity. There was little correlation between phage typing and restriction enzyme activity, suggesting that restriction and modification are not responsible for phage specificity among entomopathogenic B. sphaericus strains.     

Various levels of cross-resistance to Bacillus sphaericus strains in Culex pipiens (Diptera: Culicidae) colonies resistant to B. sphaericus strain 2362.

We studied the cross-resistance to three highly toxic Bacillus sphaericus strains, IAB-59 (serotype H6), IAB-881 (serotype H3), and IAB-872 (serotype H48), of four colonies of the Culex pipiens complex resistant to B. sphaericus 2362 and 1593, both of which are serotype H5a5b strains. Two field-selected highly resistant colonies originating from India (KOCHI, 17,000-fold resistance) and France (SPHAE, 23,000-fold resistance) and a highly resistant laboratory-selected colony from California (GeoR, 36,000-fold resistance) showed strong cross-resistance to strains IAB-881 and IAB-872 but significantly weaker cross-resistance to IAB-59 (3- to 43-fold resistance). In contrast, a laboratory-selected California colony with low-level resistance (JRMM-R, 5-fold resistance) displayed similar levels of resistance (5- to 10-fold) to all of the B. sphaericus strains tested. Thus, among the mosquitocidal strains of B. sphaericus we identified a strain, IAB-59, which was toxic to several Culex colonies that were highly resistant to commercial strains 2362 and 1593. Our analysis also indicated that strain IAB-59 may possess other larvicidal factors. These results could have important implications for the development of resistance management strategies for area-wide mosquito control programs based on the use of B. sphaericus preparations.     

Numerical classification and identification of Bacillus sphaericus including some strains pathogenic for mosquito larvae

Ninety-one strains of Bacillus sphaericus, including representatives of all the established DNA homology groups, related round-spored and oval-spored species, and six strains pathogenic for mosquito larvae, were examined for 155 characters. Numerical analyses (Jaccard coefficient/average linkage clustering) based on the 88 variable features revealed 14 clusters at the 79% similarity level that contained more than one strain and 17 single member clusters. All insect pathogenic strains were recovered in a single cluster and the classification was in accord with an established classification based on DNA sequence homology. Two frequency matrices for probabilistic identification were constructed and tested. A comprehensive matrix comprising 14 mesophilic, round-spored taxa and 27 tests gave good results for identification of hypothetical median organisms, cluster overlap and identifications of representative strains (based on data generated in the classification study). Reference strains for the 14 taxa and eight additional insect pathogenic strains were examined for the 27 tests and were correctly identified with high scores using this matrix. A second matrix comprising seven taxa and 13 tests also performed well in the theoretical evaluation and correctly identified the reference strains and insect pathogenic strains.     

Identification, cloning and sequence analysis of the Bacillus sphaericus 1593 41.9 kD larvicidal toxin gene

A number of strains of the widespread aerobic soil bacterium, Bacillus sphaericus, possess crystalline inclusions of a toxin lethal to a variety of insect (larvae) which are vectors of major tropical diseases. Partial amino acid sequence data from one strain, B. sphaericus 2362 have permitted us to design oligonucleotide probes for identifying the toxin gene in the closely related B. sphaericus 1593. The gene was found to be contained within an EcoRI-HindIII fragment and was cloned in its entirety in the bacterial plasmid pUC12. The DNA sequence was determined together with the upstream and downstream controlling elements, and a sequence of 370 amino acids was deduced for the toxin protein. This is the first reported sequence of a B. sphaericus toxin gene and will facilitate further work in characterizing the genes from other strains of different virulence and host range. The data do not support the suggestion that the toxin is derived by proteolysis of a protoxin precursor.     

Cross-resistance between strains of Bacillus sphaericus but not B. thuringiensis israelensis in colonies of the mosquito Culex quinquefasciatus

Two colonies of Culex quinquefasciatus Say (Diptera: Culicidae) were selected with Bacillus sphaericus strains C3-41 and IAB59 in the laboratory for 13 and 18 generations; they attained 145,000- and 48.3-fold resistance, respectively, in comparison with a susceptible laboratory colony (SLCq) and showed very high levels of cross-resistance (8500- to 145,000-fold) to B. sphaericus strains C3-41, 1593, 2297 and 2362. They were relatively susceptible to B. sphaericus strains LP1-G and 47-6B (only 0.8- to 2.8-fold tolerance), with 24.8- to 48.3-fold cross-resistance to strain IAB59. B. sphaericus-resistant mosquito colonies remained highly susceptible to B. thuringiensis israelensis, suggesting that B.t.i. would be of value in the management of B. sphaericus-resistant Cx. quinquefasciatus colonies. The demonstration of low or no cross-resistance of two selected resistant Cx. quinquefasciatus colonies to IAB59, LP1-G and 47-6B strains of B. sphaericus and the finding of a major 49 kDa protein in these strains suggest that there is likely to be another mosquitocidal factor in the three strains.     

DNA-DNA hybridization studies and phenotypic characteristics of strains within the 'Streptococcus milleri group'

Twenty-five strains resembling 'Streptococcus milleri' were compared by DNA-DNA hybridization, by whole-cell-derived polypeptide patterns on SDS-PAGE, and by biochemical tests. Four homology groups were revealed by DNA-DNA hybridization. DNA homology groups 1, 2 and 3 were closely related and contained the type strains NCDO 2226 (Streptococcus constellatus), NCDO 2227 (Streptococcus intermedius) and NCTC 10713 (Streptococcus anginosus), respectively. DNA homology group 4 consisted of four strains received as variants of Streptococcus intermedius which were found not to be closely related to strains in groups 1-3. The data from SDS-PAGE polypeptide patterns and biochemical tests supported the recognition of three centres of variation within the 'Streptococcus milleri group' corresponding to DNA homology groups 1-3 and indicated that strains of DNA homology group 4 are members of an as yet undescribed species within the viridans streptococci.