Within-species lateral genetic transfer and the evolution of transcriptional regulation in Escherichia coli and Shigella

Background: Changes in transcriptional regulation underlie many of the phenotypic differences observed within and between species of bacteria. Lateral genetic transfer (LGT) can significantly impact the transcription factor (TF) genes which drive these transcriptional changes. Although much emphasis has been placed on LGT of intact genes, the units of transfer and recombination do not necessarily correspond to regions delineated by exact gene boundaries. Here we apply phylogenetic and network-based methods to investigate the relationship between units of lateral transfer and recombination within the Escherichia coli - Shigella clade and the topological properties of genes in the E. coli transcriptional regulatory network (TRN). Results: We carried out a systematic phylogenetic study of genetic transfer among 5282 sets of putatively orthologous genes from 27 strains belonging to the E. coli - Shigella clade. We then used these results to examine the evolutionary histories of TF genes, as well as the transcriptional regulation of lateral genes. We found evidence of LGT in 2655 (50.3%) gene sets: 678 (12.8%) show evidence of recombination breakpoints within the gene boundaries. Thus, within-and whole-gene lateral transfer is widespread among strains of E. coli and Shigella. We found that unlike global regulators, which have mostly evolved vertically, neighbour regulators (genes which regulate adjacent genes on the chromosome) have frequently been subject to transfer within the E. coli -Shigella clade. At least 56 (62%) of the 90 neighbour regulator gene sets examined show evidence of LGT, 19 (34%) of which have internal recombination breakpoints. Neighbour regulators show no evidence of co-transfer with their nearby target genes. Rather, the frequency of recombination breakpoints, and conflicting evolutionary histories among neighbour regulators and their target genes, suggest that the genomic regions encoding these genes have been constructed through successive layering of LGT events within the clade. We find no difference in the relative complexity of regulation (i.e. the number of regulators) of lateral versus vertical genes. Conclusions: Neighbour regulators show higher frequencies of transfer than other types of regulatory genes. This implicates the topological properties of regulatory genes in the TRN, and their physical proximity to targets on the chromosome, as contributing to successful LGT. The prevalence of recombination breakpoints within regulatory and target gene sets indicates that within-gene transfer has had a significant cumulative effect on the evolution of regulatory interactions in E. coli and Shigella.