Invertebrate pathogenicity and toxin-producing potential of strains of Bacillus thuringiensis endemic to Antarctica.

Several strains of Bacillus thuringiensis were previously isolated from soil in Antarctica and appeared to have physiological adaptations to this cold, nutrient-poor environment. In spite of this they could produce abnormally large, parasporal crystals under laboratory conditions. Here, they have been further characterised for toxin genes and invertebrate pathogenicity. All of the strains were positive in PCR assays for the cry1Aa and cry2 genes. This was confirmed by sequence analysis and the parasporal crystals of all strains contained polypeptides of about 130kDa. This potential for lepidopteran toxicity was borne out in bioassays of purified δ-endotoxins against larvae of Pieris brassicae: the LD(50) values of B2408 (288μg) were comparable to that of the reference strain, HD-12 (201μg). There was no activity against the nematode Caenorhabditis elegans in spite of the fact that all strains appeared to possess the cry6 gene. PCR screening for genes encoding other nematode-toxic classes of toxins (Cry5, 4 and 21) was negative. B. thuringiensis has never previously been shown to be toxic to Collembola (springtails) but the purified δ-endotoxins of one of the Antarctic strains showed some activity against Folsomia candida and Seira domestica (224μg and 238μg, respectively). It seems unlikely that the level of toxicity demonstrated against springtails would support a pathogenic life-style in nature. All of the strains were positive for genes encoding Bacillus cereus-type enterotoxins. In the absence of higher insects and mammals the ecological value of retaining the toxic capability demonstrated here is uncertain.