The genus Bacillus consists of a heterogeneous group of Gram-positive heterotrophic aerobic or facultative anaerobic bacilli with the ability to form environmentally resistant, metabolically inert spores . These soil-borne organisms are ubiquitous throughout the world, and occupy surprisingly diverse environments [2, 3]. Within this large genus, the B. cereus sensu lato group consists of six species [B. anthracis (Ba), B. cereus (Bc), B. mycoides, B. pseudomycoides, B. thuringiensis (Bt), and B. weihenstephanensis], based on classical microbial taxonomy . However, newer molecular phylogenies and comparative genome sequencing suggests that these organisms should be classified as a single species . On the surface, this conclusion seems difficult to reconcile with the varied biological characteristics of these organisms. Some Bc strains are thermophiles , while B. weihenstephanensis is psychrophilic . By contrast, many members of this group are mesophiles, and can be found in a variety of locales including soil, on plant surfaces and in the mammalian gastrointestinal microflora . Some members of this group appear to be nonpathogenic, while others cause diverse diseases including gastroenteritis, food poisoning , endophthalmitis , tissue abscesses [10, 11], and anthrax . Bt strains have the capacity to cause disease in insects [12, 13] and possibly nematodes [14–16], while some evidence suggests that Bc strains are part of the normal insect gut flora [8, 17]. Nevertheless, whole genome comparisons between these organisms reveal a surprising similarity in gene content, and Han et al.  have concluded "that differential regulation [of gene content] modulates virulence rather than simple acquisition of virulence factor genes", a conclusion confirmed by other studies . Consequently, we will refer to these organisms as the Bc species-group, to reflect the extremely close phylogenetic relationships between these organisms.
How do we account for the underlying genomic similarity of these phenotypically diverse microbes? It has been established for some time that the most rapidly evolving and evolutionarily flexible portions of the bacterial genome are regulatory sequences and transcriptional networks [20–22]. Thus, it is no surprise that major differences between Bc species-group organisms reside in the regulation of gene expression rather than gene content. A prime example of this divergence is the PlcR-PapR quorum-sensing operon, present in all Bc species-group organisms, but harboring point mutations that differentiate group members from one another [23, 24]. The papR locus encodes a quorum-sensing signal (a secreted peptide) that is internalized and binds to PlcR, a transcriptional activator that controls gene expression and is important for Bc virulence. There are four distinct phylogenetic groups of the PapR peptide, each with point mutations that result in a unique quorum-sensing 'pherotype' . The PlcR sensor in each pherotype has co-evolved to exclusively bind only its cognate PapR peptide, and each PlcR pherotype is consequently 'blind' to the quorum sensing signals secreted by other Bc pherotypes. Ba strains (and a low percentage of Bc strains)  have taken PlcR-PapR divergence a step further. These organisms carry a unique nonsense mutation in PlcR that inactivates the quorum-sensing function entirely. Since PlcR and the global virulence regulator AtxA on the virulence plasmid pXO1 appear to antagonize one another , PlcR inactivation after Ba acquired pXO1 appears necessary for full virulence of Ba.
This is not to say that horizontal gene transfer and genome reduction have not been important in remodeling genomes within the Bc species-group. For instance, the virulence plasmids pXO1 and pXO2 in Ba appear to have been acquired by horizontal gene transfer , and represent 52% of the unique coding capacity found in the Ba genome. Although these genes have a significant impact on the Ba pathogenic phenotype, this plasmid gene content comprises only 176 genes, representing a small fraction of the total coding capacity of the Ba genome. Genome reduction has played a modest role in divergence of the Bc species-group , likely being responsible for the reduced genome size of Bc NVH391-98. However, genome reduction is probably more important for speciation events; e.g., the M. leprae genome is fully 26% smaller than that of M. tuberculosis, and carries over 1100 pseudogenes with functional orthologs in M. tuberculosis. GR has essentially eliminated 50% of the coding capacity of the M. leprae genome . Thus, subtler genome alterations within the Bc species-group, such as gene duplication, divergence and point mutations probably have contributed as much or more than horizontal gene transfer and genome reduction to the unique niche adaptations of individuals within the Bc species-group.
Anderson et al.  first noted that the genomes of Bc species-group organisms appeared to harbor an overabundance of sigma factors, compared to B. subtilis strain 168. Bacterial sigma factors bind RNA polymerase and allow the holoenzyme to recognize promoter sequences 5' to the site of initiation of transcription . Typically, bacteria encode several different sigma factors, each of which is responsible for controlling a suite of genes by activating transcription at a unique set of sigma factor specific promoter sequences. Sigma factors generally belong to two primary categories, the sigma54 and the sigma70 families . The sigma54 proteins encoded by the Bc species-group are very highly conserved, and ubiquitously present as a single copy gene. Therefore, a phylogenetic analysis of these proteins in the Bc species-group was not particularly revealing (data not shown). We consequently focused further efforts on the sigma70 proteins. Sigma70 proteins can be further differentiated into primary alternative (PA) sigma factors and extracytoplasmic function sigma factors (ECF) . In general, PA sigma factors control expression of many housekeeping functions of the cell (e.g., B. subtilis SigA), and allow the organism to respond to specific environmental stimuli such as heat-shock (e.g., SigB) [31, 32]; in B. subtilis, several PA sigma factors are integral to the sporulation developmental pathway [33, 34]. ECF sigma factors typically activate gene expression in response to extracellular signals such as the availability of specific iron sources [35, 36] and commonly are essential for disease pathogenesis [37–39]. The activity of a PA or (more commonly) an ECF sigma is often controlled by an anti-sigma factor that renders the sigma factor in a state unable to bind RNA polymerase. Activation of the sigma factor for RNA polymerase binding and transcription initiation is triggered by a signal (ligand binding, covalent modification or proteolysis) that inactivates the anti-sigma factor .
Thus, sigma factors activate transcription in response to environmental or developmental signals, and selectively activate transcription by recognizing different consensus promoter sequences to tailor gene expression to those signals . This suggested to us that many of the phenotypic differences between members of the Bc species-group organisms might be a consequence of the sigma factor gene expansion , accompanied by divergence among the sigma factor regulons of these organisms. Consequently, we began to explore the phylogeny of the sigma factors found in various Bc species-group members, by comparison to the experimentally well-understood model organism B. subtilis. To place these studies in context, we began by constructing a phylogeny of the Bacillaceae using whole-genome single copy genes. This phylogeny suggested that the current taxonomic affiliation of many members of the Bacillaceae should be reconsidered. Using this phylogeny as a basis, we then examined the phylogenetic relationships of the sigma factors encoded by members of the Bc species-group. We find that the overabundance of sigma factors encoded by the Bc species-group organisms is specifically in the ECF sigma factors, rather than in the sigma factor group as a whole. The sigma factor gene family encoded by the Bc species-group is the end-product of a dynamic gene-duplication and gene-loss process that has, until now, underestimated the true heterogeneity of ECF sigma factor content in the Bc species-group. Further, the sigma factor content carried by any given member of the Bc species-group suggests that both shared and unique gene expression patterns have evolved during the divergence of this group of organisms from a common ancestor.