Bacterial Factors Associated with Lethal Outcome of Enteropathogenic Escherichia coli Infection: Genomic Case-Control Studies

BACKGROUND Typical enteropathogenic Escherichia coli (tEPEC) strains were associated with mortality in the Global Enteric Multicenter Study (GEMS). Genetic differences in tEPEC strains could underlie some of the variability in clinical outcome. METHODS We produced draft genome sequences of all available tEPEC strains from GEMS lethal infections (LIs) and of closely matched EPEC strains from GEMS subjects with non-lethal symptomatic infections (NSIs) and asymptomatic infections (AIs) to identify gene clusters (potential protein encoding sequences sharing ≥90% nucleotide sequence identity) associated with lethality. RESULTS Among 14,412 gene clusters identified, the presence or absence of 392 was associated with clinical outcome. As expected, more gene clusters were associated with LI versus AI than LI versus NSI. The gene clusters more prevalent in strains from LI than those from NSI and AI included those encoding proteins involved in O-antigen biogenesis, while clusters encoding type 3 secretion effectors EspJ and OspB were among those more prevalent in strains from non-lethal infections. One gene cluster encoding a variant of an NleG ubiquitin ligase was associated with LI versus AI, while two other nleG clusters had the opposite association. Similar associations were found for two nleG gene clusters in an additional, larger sample of NSI and AI GEMS strains. CONCLUSIONS Particular genes are associated with lethal tEPEC infections. Further study of these factors holds potential to unravel the mechanisms underlying severe disease and to prevent adverse outcomes.