For centuries scientists have been trying to figure out why some bacteria cause disease, while others are relatively harmless or even beneficial. The search has usually focused on identifying the genetic basis for the ability to cause disease, whether it be the mutation of an existing gene or the gain of a gene via horizontal transfer.  The latter process is (mostly) limited to bacteria, where gene sharing can enable niche exploration and adaptation.

Of special interest is the Bacillus cereus sensu lato species, which contains lineages of medical and agricultural importance: B. cereus, known to cause food poisoning; B. anthracis, which causes anthrax; and B. thuringiensis, an insect and nematode pathogen widely used in agriculture (and whose Bt toxin is engineered into Bt corn and cotton). We've known for decades that the main difference between these three lineages is the presence of particular plasmids encoding proteins that are toxic to humans (B. anthracis) or invertebrates (B. thuringiensis). When I was a graduate student delving into the world of Bacillus evolution, it seemed relatively cut and dry: each lineage had its own complement of plasmids that enabled existence in its particular niche.  But as is usually the case in science, the straightforward story becomes more complicated.  In this case, additional strain sampling and gathering of genome data told us that sometimes B. cereus strains carried plasmids similar to those found in B. anthracis, and that these lineages are not necessarily on separate evolutionary trajectories.

Does this mean that a B. cereus strain could easily gain the ability to produce anthrax toxin? Or was there something that predisposed the B. anthracis ancestral lineages to pathogenesis upon gain of the toxin-producing plasmid? A recent study by Zwick et al. (http://www.ncbi.nlm.nih.gov/pubmed/22645259) used genome sequences from over 50 B. cereus sensu lato strains to address these questions. The authors examined gene content, recombination rates, strain relationships, and adaptive changes to address whether there was anything unique to the B. anthracis lineages (besides the plasmids) that could explain the adoption of a virulent lifestyle.  For example, perhaps the gain of an additional gene on the chromosome, recombination with another pathogenic lineage, or positive selection of an adaptive mutation pre-adapted B. anthracis lineages to mammalian pathogenesis. Once pre-adapted, the bacteria need only pick up the toxin-producing plasmid to be a full-fledged pathogen. The punchline of Zwick et al's study: no evidence for pre-adaptation.  What this means is that, besides the toxin-encoding plasmids, there are no genetic changes that distinguish the B. anthracis strains from B. cereus and B. thuringiensis strains.

The implications of this: to become pathogenic like B. anthracis, other B. cereus sensu lato lineages simply need to obtain the toxin-producing plasmid (and the second plasmid that encodes the ability to hide from the immune system; I haven't mentioned this here). What is the probability of this happening?  Would that mean a greater abundance of anthrax-causing bacteria? And, what are the implications for the agricultural application of B. thuringiensis for insect biocontrol?

Although the results from this paper, and many others studying this fascinating group of bacteria, could easily be turned into a doomsday scenario, I don't feel too worried. But I do think we should learn more about their evolution and ecology. Such knowledge could help prevent human disease and facilitate the discovery of additional pathogens of agricultural or biomedical interest.