The establishment of O2 gradients in liquid columns by bacterial metabolic activity produces a spatially-structured environment. This produces a high-O2 region at the top that represents an un-occupied niche which could be colonised by biofilm-competent strains. We have used this to develop an experimental model system using soil-wash inocula and a serial-transfer approach to investigate changes in community-based biofilm-formation and productivity. This involved ten transfers of mixed-community or biofilm-only samples over a total of 10–60 days incubation. In all final-transfer communities the ability to form biofilms was retained, though in longer incubations the build-up of toxic metabolites limited productivity. Measurements of microcosm productivity, biofilm-strength and attachment levels were used to assess community-aggregated traits which showed changes at both the community and individual-strain levels. Final-transfer communities were stratified with strains demonstrating a plastic phenotype when migrating between the high and low-O2 regions. The majority of community productivity came from the O2-depleted region rather than the top of the liquid column. This model system illustrates the complexity we expect to see in natural biofilm-forming communities. The connection between biofilms and the liquid column seen here has important implications for how these structures form and respond to selective pressure.

<p>Static incubation of liquid microcosms results in a physically heterogeneous environment, where depletion of O<sub>2 </sub>in the lower region creates a relatively high-O<sub>2 </sub>niche directly below the air-liquid (A-L) interface. This has been investigated using the model bacterium <em>Pseudomonas fluorescens</em> SBW25 and the biofilm-forming adaptive mutant known as the Wrinkly Spreader. In this system, colonisation of the A-L interface by the Wrinkly Spreader provides a fitness advantage over non-biofilm-forming competitors, including the ancestral SBW25, due to better access to O<sub>2</sub> in an otherwise O<sub>2</sub>-growth limiting environment. Our current research seeks to understand how the ecological interactions of this simple system applies in more complex communities, where biofilms can be produced by multiple competing or co-operative strains and the low-O<sub>2</sub> region colonised by a range of strains capable of micro-aerobic growth. Here we report the effect of selection on the productivity of A-L interface biofilm-forming communities initiated by soil-wash (SW) inocula, which were serially transferred across ten microcosms and sixty days with mixed-community or biofilm-only samples. Initial analysis of the serial transfer experiments shows a decrease in community productivity which is explained by the accumulation of toxic metabolites, though small increases in community biofilm strength and attachment were also observed. Isolate-level analysis revealed a decrease in community diversity and a biofilm-associated phenotypic shift between the SW inocula and final-transfer communities, and these changes provide evidence of selection within our system.</p> <p>Cell-localisation experiments confirm enrichment at the top of the liquid column in the high-O<sub>2</sub> region, but also show high cell densities in the low-O<sub>2</sub> region, even within the biofilm-only final-transfer communities. Samples taken from the biofilm and lower region of these communities were able to re-colonise both in fresh microcosms, indicating that community members were capable of migration within the liquid column. Despite the over-all decrease seen in community productivity in the serial transfer experiments, we suggest that communities maximised productivity by colonising both regions of the liquid column, with a resource trade-off between fast growth in the highly competitive high-O<sub>2 </sub>region and slower growth in the less-competitive low-O<sub>2</sub> region. Many isolates from the final-transfer communities could occupy both regions and were capable of migration, with almost all isolates capable of flagella-mediated motility, and we interpret this ability to move between regions as a fitness advantage in A-L interface biofilm-forming communities. Although we have not been able to test this directly using the final-transfer communities or isolates, we have been able to demonstrate a fitness advantage in the less complex <em>P. fluorescens</em> SBW25 system, where biofilm-forming mutants capable of colonising both regions had a greater competitive fitness advantage over those with a poor ability to colonise the liquid column.</p>

Static microcosms are a well-established system used to study the adaptive radiation of Pseudomonas fluorescens SBW25 and the adaptive biofilm-forming mutants known as the Wrinkly Spreaders (WS). We have developed this system to investigate selection within multi-species communities using a soil-wash inoculum dominated by biofilm-competent pseudomonads. Here we present community and isolate-level analyses of one serial-transfer experiment in which replicate populations were selected for over ten transfers and 60 days. Although no significant trends in improving community biofilm characteristics or total microcosm productivity were observed, a significant shift in biofilm-formation and microcosm growth by individual isolates recovered from the initial soil-wash inoculum and final transfers indicated that these communities were subject to selection for growth in these microcosms. Surprisingly, the fitness of the archetypal WS was poor when competing against community samples, and having compared the cell densities in the low-O2 region of liquid column below the biofilm, we suggest that part of the community’s fitness advantage comes from the ability to colonise this under-utilised niche as well as to compete at the A-L interface. Samples from the community biofilms and the low-O2 region were able to re-colonize both niches and many final transfer isolates grew throughout the liquid column as well as forming A-L interface biofilms. This suggests that there is a trade-off between fast growth under highly competitive conditions at the A-L interface and slower growth with less competition in the low-O2 region, with some isolates taking a bet-hedging approach a colonizing both niches in our microcosm system.