On average,

the dispersal isolates of strain 18A gained or lost the ability to utilise four substrates, where the greatest gain of function was four (18AWT-1 and -3) and the greatest loss was six (18ASTY-5, Table 2). Of the morphotypically different, biofilm-derived isolates, one isolate, 18ASTY-1, had the same profile as isolate 18AWT-10. The remaining nine 18ASTY variants were classified into five novel profiles (profiles 7–11, Table 2). The 18AWT and 18ASTY biofilm-derived isolates commonly gained the capacity to utilise α-keto butyric acid and find more 2, 3-butanediol and most frequently lost the ability to use d-alanine, l-ornithine d-trehalose. In contrast to the variable substrate utilisation observed for the wild type (WT) 18A dispersal variants, all of the WT PAO1 dispersal isolates shared the same metabolic profile as the parental PAO1. However, with the exception of the PAO1SCV-2 and PAO1SCV-6, the SCVs derived from PAO1 differed in their substrate utilisation patterns from PAO1 and were grouped into seven different profiles (Table 3). PAO1SCV-1 gained the capacity to use 12 substrates, which was the greatest change observed for any of the isolates

tested. Interestingly, two PAO1 SCVs (PAO1SCV-1, -5) gained the ability Rucaparib in vivo to grow on α-keto butyric acid and three lost the ability to grow on 2, 3-butanediol (PAO1SCV-4, -5, -7). As noted above, these substrates were also ones for which utilisation was altered in some of the 18AWT and 18ASTY dispersal cells. However, for the PAO1SCVs, the ability to utilise 2, 3-butanediol

was the most commonly lost, whilst it was most commonly gained in the strain 18A variants. As an additional (-)-p-Bromotetramisole Oxalate control, 10 isolates each from an overnight culture of strains 18A and PAO1 with the WT morphotype were tested for their substrate utilisation patterns and were found to be identical to their respective parents (data not shown). Therefore, it appears that phenotypic variation, as determined here, is enhanced during biofilm growth and dispersal. Biofilm-derived dispersal isolates of strain 18A (18AWT and 18ASTY) were compared with the parental 18A strain for attachment and biofilm formation on hydrophobic and hydrophilic surfaces. Similar results were obtained for both surfaces, and hence, only the data for the hydrophobic surfaces are presented (Fig. 2). Overall, extensive variability was observed in the attachment (Fig. 2a) and biofilm formation (Fig. 2b) for all of the dispersal isolates of 18A (WT and STY). While PAO1 biofilm-derived isolates also showed considerable variation in attachment and biofilm formation (Fig. 2c and d), the overall variability was less than that observed for the 18A biofilm-derived variants.