Fig. 2. Putative gradients of nutrients or cell products do not explain movement towards antibiotics.
A Cells move towards increasing [tryptone] (black line, CMAX = 100% of the concentration used in growth medium, CMIN = 10%; a two-sided binomial test at t = 5 h rejects the null hypothesis that trajectories are equally likely to be directed towards or away from tryptone (p < 0.0001, n = 132 cell trajectories). However, when ciprofloxacin (CMAX = 10X MIC) is added to the lower inlet, cells initially (t < ≈7.5 h) move towards increasing [tryptone] (blue line; p < 0.0001 at t = 5 h, n = 110), but then the bias is reversed towards increasing [ciprofloxacin] (p < 0.0001 at t = 15 h, n = 1219). B Cells move away from cell-free supernatant (purple line, CMAX = 10%; p < 0.0001 at t = 9 h, n = 1464). However, when ciprofloxacin (CMAX = 10X MIC) is added to the supernatant, cells rapidly (after ≈ 5 h) move towards increasing [ciprofloxacin] (and thus increasing [supernatant]; light-green line, p < 0.0001 at t = 13 h, n = 1059). Cells similarly move towards ciprofloxacin in a uniform background concentration of 10% cell-free supernatant (dark-green line, p < 0.0001 at t = 13 h, n = 583). C After ≈20 h, YFP-labelled WT cells in a ciprofloxacin gradient (CMAX = 10X MIC) form dense biofilm at [ciprofloxacin] <1X MIC, whilst a smaller band of migrating cells is visible at much higher concentrations. D Coloured regions showing the antibiotic gradient (∂C/∂y) magnitude using our standard flow speed. E, F Increasing flow threefold sharpens the gradients, allowing both the 1X MIC isocontour (dashed white line) and the band of migrating cells to stretch further downstream. G Movement bias, β, increases with ∂C/∂y (line colours correspond to regions shown in (F)). Least squares linear regression at t ≈ 17 h of log10(∂C/∂y) against β yielded a slope of 0.258 (95% confidence bounds = 0.191, 0.325). Figures S12 and S17 show biological repeats and source data are provided as a Source Data file.