Surface proteins of gram-positive bacteria often play a role in adherence of the bacteria to host tissue and are frequently required for virulence. A specific subgroup of extracellular proteins contains the cell wall-sorting motif LPxTG, which is the target for cleavage and covalent coupling to the peptidoglycan by enzymes called sortases. A comprehensive set of putative sortase substrates was identified by in silico analysis of 199 completely sequenced prokaryote genomes. A combination of detection methods was used, including secondary structure prediction, pattern recognition, sequence homology, and genome context information. With the hframe algorithm, putative substrates were identified that could not be detected by other methods due to errors in open reading frame calling, frameshifts, or sequencing errors. In total, 732 putative sortase substrates encoded in 49 prokaryote genomes were identified. We found striking species-specific variation for the LPxTG motif. A hidden Markov model (HMM) based on putative sortase substrates was created, which was subsequently used for the automatic detection of sortase substrates in recently completed genomes. A database was constructed, LPxTG-DB (http://bamics3.cmbi.kun.nl/sortase_substrates), containing for each genome a list of putative sortase substrates, sequence information of these substrates, the organism-specific HMMs based on the consensus sequence of the sortase recognition motif, and a graphic representation of this consensus.

Members of the Bacillus cereus group are ubiquitously present in the environment and can adapt to a wide range of environmental fluctuations. In bacteria, these adaptive responses are generally mediated by two-component signal transduction systems (TCSs), which consist of a histidine kinase (HK) and its cognate response regulator (RR). With the use of in silico techniques, a complete set of HKs and RRs was recovered from eight completely sequenced B. cereus group genomes. By applying a bidirectional best-hits method combined with gene neighbourhood analysis, a footprint of these proteins was made. Around 40 HK-RR gene pairs were detected in each member of the B. cereus group. In addition, each member contained many HK and RR genes not encoded in pairs (‘orphans’). Classification of HKs and RRs based on their enzymic domains together with the analysis of two neighbour-joining trees of these domains revealed putative interaction partners for most of the ‘orphans’. Putative biological functions, including involvement in virulence and host–microbe interactions, were predicted for the B. cereus group HKs and RRs by comparing them with those of B. subtilis and other micro-organisms. Remarkably, B. anthracis appeared to lack specific HKs and RRs and was found to contain many truncated, putatively non-functional, HK and RR genes. It is hypothesized that specialization of B. anthracis as a pathogen could have reduced the range of environmental stimuli to which it is exposed. This may have rendered some of its TCSs obsolete, ultimately resulting in the deletion of some HK and RR genes.