15 mL min−1 for each channel) at room temperature as previously described (Moller et al., 1996). Bacteria

were inoculated into 10 mL of LB10 broth and incubated overnight (16–18 h) at 37 °C with shaking. Flow-cell channels were inoculated with 5 mL of the overnight culture and incubated without flow for 1 h for PAO1 and 2 h for 18A at room temperature, owing to the decreased efficiency of attachment for 18A, as described by O’May et al. (2006). M9 medium containing 48 mM Na2HPO4, 22 mM KH2PO4, 9 mM NaCl, 19 mM NH4Cl, 2 mM MgSO4, 100 μM CaCl2 and 5.5 mM glucose was used to cultivate flow-cell biofilms. All chemicals were purchased from Univar, Australia, unless otherwise indicated. Every 2 days, 2 mL of biofilm effluent was collected from the outflow of the biofilm flow cell, serially diluted using M9 salts solution without glucose and spread-plated onto LB10 agar plates. CFUs RO4929097 cell line JQ1 and biofilm variants were enumerated according to the colony morphologies exhibited after 2 days of incubation at 37 °C. As a control, planktonic cultures of the parental strains were inoculated into 10 mL of M9 medium and cultured at 37 °C with shaking and subcultured daily after overnight growth for 14 days. Sampling was performed every

2 days, and these cultures were serially diluted using M9 salt solution and spread-plated onto LB10 agar plates to detect and quantify phenotypic variants. For phenotypic characterisation, variants from PAO1 and 18A biofilms were collected PRKACG on days 6 and 10, respectively, which were determined by confocal laser scanning microscopy to correlate with hollow colony formation and cell death, hallmarks of dispersal. The mutation frequency of strains 18A and PAO1 was quantified as described by Oliver et al. (2002). Briefly, independent triplicate cultures were grown in LB10 broth (overnight, with agitation) and serially diluted, and 100-μl aliquots were plated onto LB10 agar or LB10 agar containing rifampicin (300 μg mL−1). Plates

were incubated at 37 °C for 48 h. Mutation frequencies were estimated as the mean number of rifampicin-resistant CFU divided by the total CFU (LB10 agar plates without rifampicin). Biofilms of both PAO1 and 18A strains were cultivated as continuous cultures in silicon tubing (Silastic® Laboratory Tubings) as described by Barraud et al. (2009). Briefly, 5 mL of the overnight culture was inoculated into each tube (inner diameter 2.64 mm) using a syringe, under conditions of no flow, and the injection site was subsequently sealed with silicone glue (Plastic Putty; Selleys Pty Ltd, Australia). The inoculated tubes were incubated under static conditions for 1 or 2 h for strain PAO1 and strain 18A, respectively, to allow bacteria to attach to the walls of the tubing, after which time medium flow was resumed at a rate of 0.15 mL min−1.