The evolution of silent beta-glucoside systems in Escherichia coli

The genomes of bacterial species are very dynamic. For example in Escherichia coli, individual isolates may share as little as 60 to 70% of their genome with other isolates. The remainder of the genome consists of a flexible pool of genes, which are present only in some isolates. This genome diversity is manifested through gain of genes by horizontal transfer as well as by loss or mutations of genes. In this study the evolution of two silent β-glucoside loci belonging to the flexible gene pool of E. coli was traced. The bgl operon is present in ~80% of all E. coli isolates, while in the rest it is replaced by a locus (named Z5211-5214) of unknown function. The bgc locus is present in roughly 50% of E. coli isolates. To trace the evolution of these loci, the phylogeny of a representative collection of 175 E. coli isolates was established by multilocus sequence typing (MLST). In parallel, the bgl and bgc loci were typed by PCR and sequencing. This revealed four groups of the bgl / Z locus, including 3 bgl variants. Mapping of these groups demonstrated a striking correlation to the species phylogeny and population structure: among the four phylogenetic groups of E. coli, bgl is present in the A, B1, and B2 groups, while the Z locus is present in D strains, which suggests a coupled evolution of the bgl / Z locus with the host. Further, three ancestral E. coli isolates and strains of the closely related species Escherichia albertii as well as the closely related genus Salmonella enterica, lack the bgl / Z locus, indicating horizontal transfer of the bgl and Z loci into the root of the modern E. coli. However, BLAST surveys revealed the presence of bgl homologs in Erwinia, Klebsiella and Photorhabdus species. The phylogeny of E. coli bgl and these homologs is concordant with the 16S rRNA phylogeny contradicting horizontal transfer. In conclusion, these results implicate vertical inheritance of bgl and its loss in some enterobacterial lineages. In E. coli isolates belonging to the phylogenetic group D, the bgl operon presumably was replaced by the Z locus, which may have been horizontally acquired. Further, correlating the data of a functional analysis of bgl with the species phylogeny demonstrated that bgl is functional although silent in the majority of strains in groups A and B1, while, interestingly, in more than 50% of B2 strains, bgl was not silent but weakly expressed. These data together with the results of nonsynonymous-to-synonymous substitution ratio test (the KA/KS test), suggest that bgl may confer an unknown ecological advantage. The second silent β-glucoside system bgc analyzed here, is predominant in the phylogenetic groups B1 and B2, it is present in D group isolates and rarely found in A strains. The widespread occurrence could be due to either gain or loss of bgc in evolution. Cumulatively, the study suggests that in addition to gene gain, also gene loss may significantly contribute to the flexibility of the genome, and that a careful analysis is required for individual loci belonging to the flexible gene pool