Fig. 4: Dose-response curves to ciprofloxacin (CIP) of E. coli MG1655 (wildtype, WT), the BAC-tolerant S4 strain and their ompF knockouts.

The y-axis shows the growth rate of the strains relative to the growth rate of the WT without CIP. The x-axis shows the CIP concentration on a logarithmic scale. The WT is represented in brown color, S4 in yellow, S4ΔompF in pink and WTΔompF in blue. The lines show a logistic fit of the data, fitting 3 parameters: upper limit, the half-maximal inhibitory concentration (EC₅₀) and slope. The lower limit was fixed to 0. The shaded area represents the 95 % confidence interval according to the fit. The model showed a strong predictive potential across all strains with R² = 0.96 and L² = 0.998, indicating that no linear corrections are needed to fulfill its predictive potential. The predictive potentials for each curve individually are as follows: R² = 0.99 and L² = 0.999 for WT, R² = 0.96 and L² = 0.997 for S4; R² = 0.90 and L² = 0.98 for S4∆ompF; R² = 0.92 and L² = 0.99 for WT∆ompF. The MIC values of the used strains as measured here are WT: 0.005–0.01 µg∙mL⁻¹(median=0.0075), S4: 0.04–0.08 µg∙mL⁻¹(median=0.06), WTΔompF: 0.16–0.18 µg∙mL⁻¹(median=0.17), S4ΔompF: 0.12–0.16 µg∙mL⁻¹(median=0.14).