Abstract
"Avian Pathogenic Escherichia coli (APEC) cause extra-intestinal infections in birds, leading to significant welfare issues and economic losses. Multiple lineages of APEC are associated with disease in poultry, and complicate intervention strategies such as vaccines. Genome scale metabolic models (GEMs) are a powerful tool for the analysis and prediction of metabolic capability and gene essentiality. Limited number of GEMs are currently available for E. coli (including APEC). Therefore, this thesis aimed to improve the understanding of APEC, and construct and validate a pan genome-based APEC GEM.
In this study, a diverse panel of APEC isolates were genotypically and phenotypically characterised. 114 representative APEC isolates were genome sequenced and used for model construction. Knockout mutants of representative APEC isolates were constructed for metabolic model validation.
APEC and chicken commensal isolates showed significant differences in phylogroup distribution (P<0.005), and O78 (17%) and O2 (13%) serotypes were predominant in APEC, and mainly clustered in phylogenetic groups C and B2, respectively. No single or combination of virulence factors were representative for APEC.
A pan-genome based GEM representing the 114 APEC isolates (APEC GEM) contained 2923 reactions, 2527 metabolites and 2242 genes. The correlation coefficient between computational prediction and experimental observation of three representative APEC isolates (from phylogroup B2, C and G, respectively) indicated that under same constrained nutrient condition, APEC GEM showed a higher agreement to our APEC panel compared to the E. coli K12 MG1655 GEM. The model accurately predicted lysine as an essential pathway for APEC and a lysA mutant was observed to be auxotrophic for L-lysine. Phylogroup-specific sub-models derived from APEC GEM demonstrated significant potential for distinguishing metabolic profiles between different genotypes of APEC.
The studies presented here demonstrate that APEC are a diverse pathotype. The APEC GEM described here can serve as draft for further construction of genotype specific sub-models."