A genomic population genetics analysis of the pathogenic enterocyte effacement island in Escherichia coli: the search for the unit of selection

Abstract

Comparative genomic analysis is a powerful tool for understanding the history and organization of complete genomes. The mathematical tools of population genetics combined with genomic analysis provide a powerful approach to dissect heterogeneities in genome evolution. This study presents a hierarchical analysis of the enterocyte and effacement island (35 kb), which is found in the enteropathogenic and enterohemorrhagic strains in Escherichia coli and in Citrobacter rodentium. The locus of enterocyte and effacement in E. coli is considered to be a clonal unit inside a clonal organism and is expected to evolve as a single unit. This analysis examines the clonal assumption by determining genetic diversity, GC content, and the substitution rates at the different functional levels of (i) the complete pathogenic island, (ii) the five operons in which the island is organized, and (iii) for each of the individual 41 genes that comprise the locus. We find that there is a conserved region that is composed of genes that belong to the type III secretion system and that may be products of horizontal transfer. A more diverse region is composed of genes for secreted proteins and genes that we infer to be original components of the E. coli genome. This genetic mosaic seems to be differentially affected by selection and mutation. Our results suggest that recombination and selection may be breaking this structure so that different elements are, at best, weakly coupled in their evolution. These observations suggest that the units of selection are not the complete island, but rather, much smaller units that comprise the island.

Fig. 1.

Genealogy of the six LEE islands used in the study. The genealogy was constructed by using the 32,148 bp that comprise the core consensus of the six islands (including coding and noncoding sites), under the neighbor-joining method with Tamura–Nei distance, and 10,000 bootstrap resamplings. STEC, Shiga toxigenic E. coli.

Fig. 2.

GC content, genetic diversity, and dN/dS ratio distribution for the 41 genes of LEE.

Fig. 3.

Genealogy of nine LEE genes used in the study. The genealogy of each gene was constructed under the neighbor-joining method with Tamura–Nei distance and 5,000 bootstrap resamplings. The genes escJ, escV, escN, and escF belong to the TTSS. The genes espH, map, and espA are secreted proteins, and tir is the receptor for the adhesin eae.

Fig. 4.

Split-decomposition analysis of eight LEE genes.