Genome-wide fitness analyses of the foodborne pathogen Campylobacter jejuni in in vitro and in vivo models: whole-genome sequence analysis of defined gene deletion mutants.

Here, we report a detailed analysis of C. jejuni fitness across models reflecting stages in its life cycle. Transposon (Tn) gene-inactivation libraries were generated in C. jejuni and the impact on fitness during in vitro growth, chicken colonisation, survival in houseflies and under nutrient-rich and –poor conditions at 4 degrees and infection of human gut epithelial cells was assessed by Tn-insertion site sequencing (Tn-seq), providing a measure for the relative abundance of each Tn mutant in the library. Genes identified to be required under the tested conditions were validated using a panel of define gene deletion mutants. C. jejuni M1cam gene deletion mutants were constructed by allelic replacement of the gene with a chloramphenicol (cat) resistance cassette. Gene deletion mutants were subjected to phenotypic and genotypic characterisation, including motility and in vitro growth in liquid culture. Whole genome sequencing (WGS) based variant analysis (single nucleotide polymorphisms [SNP] and insertion/deletions [INDELS]) of defined deletion mutants was used to screen for second-site mutations that might have affected the phenotypes under investigation. Collectively, variants were detected at 129 positions relative to the C. jejuni M1cam reference genome. The variant database was cross-referenced with data obtained in Tn mutant library screens to assess whether the gene affected by the variant had a potential impact on the phenotype under investigation. In addition, phenotypes of deletion mutants sharing a variant were compared to predict possible confounding effects of the second-site mutation. This in-depth analysis did not identify confounding effects of second-site mutations. Sequencing libraries were prepared using the NEBNext Ultra or Ultra II DNA library prep kit (New England Biolabs) and sequenced on the MiSeq platform. Reads were mapped to the M1cam reference genome (accession no. CP012149 13) using Stampy, variants were identified with Samtools and the effect at the protein level was predicted using SnpEff.