Fig. 3. Cells migrate towards antibiotics and die.
A YFP-labelled WT cells (yellow) and unlabelled, chemotaxis-impaired mutant cells (ΔpilG, not visible in this epi-fluorescent image) were exposed to ciprofloxacin gradients (CMAX = 100X MIC) in fluid-walled microfluidic devices. At t ≈ 24 h, channels were reconfigured (dashed red lines) in separate experiments at either ‘Position A’, isolating all chemotaxing cells, or ‘Position B’ or ‘C’, which were progressively more conservative and isolated cells that had migrated further towards CMAX. Cartoon B and experimental image C corresponding to the approximate location of the dashed-cyan box in (A) immediately prior to channel reconfiguration. We predicted – and experimentally confirmed – that the WT (yellow), but not ΔpilG (cyan), would move towards ciprofloxacin. Image representative of four independent experiments. D Channel reconfiguration aimed to isolate chemotaxing WT cells in an isolated microfluidic chamber containing antibiotic-free nutrient medium. E Experimental image of the cells after channel reconfiguration confirms formation of an isolated fluid chamber with mostly WT cells (mean proportion of WT over nine independent samples = 84% with 95% confidence intervals of 81-87%; a two-sided Z-test rejected the null hypothesis of 50% WT cells, p < 0.0001, n = 5074). F Planktonic cells extracted from the chamber and monitored for growth on nutrient agar showed no detectable growth. G Chambers were imaged for ≈40 h to monitor any remaining cells. No cell growth was detectable when channels were reconfigured at Positions B and C and the chamber surfaces gradually cleared as cells died or detached (see also the image series marked by †, which is representative of the three separate experiments). When reconfigured at Position A, the fraction of the chamber surface covered by cells initially decreased, momentarily increased, and finally plateaued at ≈10% coverage. After t ≈ 60 h (see*), we extracted the entire chamber contents and monitored growth on agar plates lacking antibiotics. H After a further 24 h, we recovered a small number of both WT and ΔpilG colonies in the Position A experiment, suggesting we had isolated some cells that had not yet reached lethal antibiotic concentrations. Source data are provided as a Source Data file.