Accumulation of colibactin-induced mutations in E. coli

A lab E. coli strain was repeatedly exposed to colibactin and a control strain to evaluate mutations induced by colibactin. An ampicillin-resistant BW25113 reporter strain was grown overnight in LB with 50 ug/mL carbenicillin. 1 mL of the overnight was washed three times in PBS before diluting 1:50 in M9 media and growing for an additional two hours at 37C. OD600 was measured and the culture was diluted to OD600=0.1. The BW25113 culture was mixed 10:1 (toxic to reporter) with either a colibactin or non-colibactin producing strain (cultured overnight, washed, grown for two hours, and diluted to OD600=0.1 as the reporter was) to a final volume of 500 uL in 96 deep-well plates (Eppendorf, cat# 2231000920). The co-cultures were pelleted and incubated at 37C for 24 hours. Cultures were then mixed and 5 uL of each culture was transferred to a 96 well microplate (FisherBrand, cat# FB012932) and diluted by mixing with 95 uL of PBS. From there, each culture was serially diluted 1:5 for 7 dilutions total. 4 uL from each dilution was spotted on LB agar plates supplemented with carbenicillin to select for the ampicillin-resistant BW25113 strain and plates were incubated at 37C overnight. A single colony was picked from each replicate and inoculated into 500 uL M9 in a 96 deep-well plate. Fresh cultures of the colibactin and non-colibactin producing strains were grown overnight and the co-culture protocol was repeated. The co-culture process, from the 24-hour co-culture, dilution and spotting on selective plates, picking single colonies and growing them up, was repeated for 10 cycles. After the 10 cycles, single colonies from each replicate were expanded and frozen as glycerol stocks. Glycerol stocks were inoculated into 1 mL LB and grown for 3-4 hours before DNA was extracted with Zymo Quick DNA 96 kit (cat# D3012). Four replicates (from the same condition) were combined evenly prior to whole-genome sequencing at SeqCenter (400Mbp Illumina paired-end sequencing with 2x151bp). There were 48 replicates per condition, so after pooling four replicates for sequencing, we sequenced 24 samples in total with an average coverage of 65.