With few exceptions [9, 10] the crystal and spore of B. thuringiensis are released separately during sporulation. It was thus surprising that 24/79 non-clonal, candidate B. thuringiensis isolates possessed spores with a "capped" appearance. Some of these caps appeared to contain crystals since phase dark objects with a light center could be seen in many of the caps, including Bt2-56, which was subsequently confirmed by Electron Microscopy to possess a parasporal body within the exosporium (Figure 4). The functional significance of spore-associated crystals is not totally understood however benefits such as protection against UV degradation and better access to the target organism have been attributed to crystals attached within an exosporium . If this is true, and the caps are shown to contain useful toxins, then those isolates showing the capped morphology may eventually prove useful in field applications. Clearly, more attention needs to be given to this phenotype and its mechanism of expression. For example, if the phenotype is plasmid encoded, could strains that undesirably liberate crystals be transformed to enclose their crystals within a protective cap?
Examples of filamentous appendages associated with bacterial spores are relatively rare where only a few isolates of Clostridium and Bacillus are reported to possess spores with appendages [12, 13]. Even rarer is the association of an appendage with a parasporal body, as was observed with both Bt1-88 and Bt2-56. Very little is presently known about Bt1-88 and studies are underway to characterize both the strain and the ultra structure of the appendage-associated complex. From preliminary studies undertaken with Bt2-56, this isolate clearly demonstrates an intimate relationship between the filament and the parasporal body where the parasporal body appears to help anchor the filament to the spore (Figure 4 and ). Both these isolates raise a number of questions fundamental to our basic understanding of bacteria and the role of spore associated appendages. Among these is the obvious question of a possible relationship between the appendage and B. thuringiensis crystals, i.e., did some crystals evolve to act as an anchoring structure for spore associated filaments and if so, do they still retain any toxin activity? Similarly, what is the role of the filament and does it aid in pathogenicity?
Sequence data indicated that a gene in Bt1-33 is highly homologous to the cry14Aa1 gene, [GenBank: U13955] which has demonstrated activity against nematodes . It was surprising that this relatively small study would produce an isolate containing a putative nematocidal gene when a much larger study utilizing the same primers did not . However the gene in Bt1-33 is probably non-functional due to a frame-shift mutation, raising the question of why would it be maintained by the bacterium? One explanation could be that the gene is clustered in the genome with other genes necessary for pathogenicity as part of a pathogenicity island, and is being maintained in the genome indirectly through selection for the other genes in the group. Interestingly, PS80JJ1 (the strain gene U13955 is found in), has been shown to contain at least two other δ-endotoxin proteins Cry34Aa1 and Cry35Aa1, which together form a binary toxin shown to be active against Diabrotica virgifera (western corn rootworm). The corn root worm, like the nematode, is a soil-dwelling organism that B. thuringiensis are rarely active against. It is perhaps significant that U13955 is associated with other toxins that target a soil-dwelling organism, raising the question of whether PS80JJ1 is adapted against soil-dwelling organisms. If so, then perhaps like PS80JJ1, Bt1-33 may also possess other similarly rare toxin genes that target soil-dwelling organisms.
Using the PCR amplified fragment from the cry14Aa1-like gene in Bt1-33 as a probe, another isolate, Bt1-35, showed a weaker but definite hybridization signal. Perhaps significant is that it was isolated from the same sample as Bt1-33. The sequence of this gene is presently unknown however the reduced hybridization signal shown suggests that it may be an evolutionary relative or even share limited sequence, such as a functional domain.
The isolation method used in this study diverged from many of the reported methods for B. thuringiensis isolation by utilizing a stain instead of Phase Contrast Microscopy (which is commonly used when screening for B. thuringiensis), for the identification of crystals [1, 16, 17]. Although Phase Contrast Microscopy was useful for observing inside spore caps, the stain uniquely allowed a fast, high throughput evaluation of bacterial colonies for the presence of crystals. Similarly, the high contrast of the stain allowed it to identify relatively small crystals or low numbers of caps in a specimen. It is doubtful that without these benefits of the stain that Bt1-88, Bt1-33/35 and the relatively small crystal producers would have been isolated. This study also sought to reduce the chance of excluding some isolates by not incorporating overt selection during the isolation of candidate bacteria (e.g., not using antibiotics in the culture medium). The importance of reducing selective pressures for the successful isolation of each strain was not determined. However, at least one of the important isolates, Bt1-33, was isolated using Method G, a selective method. Thus it is possible that selective enrichment coupled with the stain and high throughput screening could have resulted in the isolation of higher numbers of B. thuringiensis candidates than were obtained in this study. Similarly, the isolation of capped isolates, including Bt2-56, could not be directly attributed to the use of the stain or even the general isolation methodology used. For example, although not as evident as with the use of the stain, Phase Contrast Microscopy could be used to visualize the capped spores. Thus it is possible that the high level of capped spores observed is best attributable to their high or unique presence in the Trinidadian environment. For example, it is difficult to imagine how the cap-associated filament of Bt2-56 has not previously been identified elsewhere, when in this screen it was isolated from several specimens collected throughout Trinidad. However, the filament was not easily observed under Phase Contrast Microscopy without some special attention being given to finding it. For example, the filament was not observed immediately with Phase Contrast Microscopy and it was only by chance that, as a capped spore, Bt2-56 was chosen for more intense study and its filament found. Similarly, the spiked structures associated with Bt1-88 were not observed on primary isolation or initial examination with Phase Contrast Microscopy, and it was only through a re-screening of all capped spores for filaments similar to Bt2-56 that it was observed. These data would suggest that some capped spores contained in existing B. thuringiensis libraries might also possess appendages and should be re-examined specifically for their presence.
Perhaps most striking about this screen was the low percentage of isolates producing bipyramidal crystals. In literally all reported screens for environmental isolates of B. thuringiensis, the percentage of bipyramidal crystals is usually above 40% [18, 1], however only 3 isolates or 3.8 % of the non-clonal isolates where shown to have a bipyramidal morphology indicating a screen that deviated significantly from the norm.