A Medically Relevant Three-Species Mixed Culture in a Chemostat : Mathematical Modelling and Experiments to Study Interaction Effects

Pseudomonas aeruginosa, Burkholderia cepacia and Staphylococcus aureus are opportunistic infectants which occur as mixed cultures in the lungs of cystic fibrosis (CF) patients. Knowledge on possible interactions and growth characteristics of the microbial community in the lung obviously cannot be obtained in situ. It would be very desirable to be able to predict the mixed culture’s reaction, e.g. on antibiotic treatment, for optimal therapy of patients. We set up a laboratory chemostat to quantitatively study the three species as a reproducible mixed culture under completely defined and controllable conditions [1]. Our focus was to find possible interactions by combination of experiments and mathematical modelling. Choice of a chemically defined culture medium primarily consisting of glucose and amino acids allowed for quantitative analysis of metabolic substrates and products. Absolute and species-specific cell numbers were determined by a quantitative T-RFLP analysis method [1]. Experiments showed an apparent coexistence of at least two of the species for more than 32 volume exchanges. This result was not predictable with a model assuming pure substrate competition [2]. Therefore, additional model assumptions were implemented and tested by comparison to experimental results. We present two selected interactions that were analysed in detail. One effect was that of a metabolic interaction, i.e. in this case the metabolic product of one species could serve as substrate for the two other species. The effect was shown mathematically to be a premise for coexistence, but not in the case for our species [2]. Another effect was that of amino acids as an additional substrate to glucose. The standard assumption for chemostat cultivations is that there is only one limiting substrate. It is known, that medium components can serve as co-substrate, mixed substrate or second carbon source [3]. This is rarely taken into account, when investigating pure or even mixed chemostat cultures. With a glucose-free variant of the chemically defined medium different hypotheses on the specific function of the amino acids were tested for each of the three species. Additionally, selected scenarios of the mixed culture outcome for variation of the implemented assumptions will be presented. We show, that even our laboratory model system is too complex to be described by the reduced assumptions of single-substrate competition, but that specific single species properties need to be considered to describe the mixed culture behaviour. The combination of mathematical modelling and quantitative experimental studies has to our knowledge not been presented before for a more than two species microbial community in such detail and quantitative quality. Further investigations will follow the same approach in studying the effects of antibiotic pulses on the dynamics of the mixed culture. [1] Schmidt, J.K. et al.: (2006) Biotech Bioeng (published online) [2] Heßeler, J. et al.: (2006) J Math Biol 53: 556-584 [3] Kovarova-Kovar K. et al.: (1998) Microbiol Mol Bio Rev: 646-66