Fundamental and applied research are sometimes thought to be mutually exclusive. However, our rather fundamental ecological study recently published in the ISME Journal started with a biotechnological question. Our initial aim was to select bacterial mutant strains that overproduce specific extracellular metabolites which act as flavour compounds in fermented products. For this we envisioned a selection-system using water-in-oil emulsions, that contain millions of separated reaction compartments in which individual bacterial cells can be grown. The product concentration inside each droplet is measured using sensor-cells that give a fluorescence signal proportional to the product concentration. As producers and sensors are co-localized in a droplet, a high fluorescence signal should allow selection/sorting of producer-cells that make most of the extracellular compound.
A prerequisite for this method to work is that sensors should only interact with producers in the same droplet. However, we quickly realized that some volatile metabolites rapidly diffused into the oil-phase of the emulsion, which disrupted this critical coupling between the product-concentration and the response of the sensor-cell. This hurdle made us study the impact of diffusion (product-leakage) on producer-sensor interactions, with the aim to establish specific producer-sensor interactions within a compartment for diffusing compounds.
To get a feeling for the shape of the concentration gradients around the producer-cells we made a reaction-diffusion model. The first experiments showed that this model could very well predict the behaviour of the interacting cells. Subsequently we tried for several months to bring producer- and receiver-cells close to each other in very small droplets but that proved difficult. As the needed interaction-distances were predicted to be in the low μm-range, we came to the conclusion that a different protocol was needed to achieve very close proximity of producers and sensor cells. Based on our model predictions and inspired by a presentation from Ales Lapanje this led to a collaboration with his group from the Jozef Stefan Institute in Slovenia. They were working on a technology to aggregate cells based on surface charge modifications, and they were happy to further develop and optimize this protocol for our strains. As soon as this method was technically up and running, in December 2018, we indeed saw that sensors could interact within such producer-sensor aggregates – a nice Christmas gift, and a beautiful illustration of mutualistic interactions between model predictions and experiments.
We realized the potential of our approach for addressing a number of questions on microbial interactions early on. Our demonstration of how competition for resources reduces interaction distances to the low um range highlights the importance of community composition and spatial structure for the evolution and maintenance of metabolic interaction. Besides its fundamental role in microbial ecology this is also a highly relevant for various biotechnological applications involving biosensors. In a follow up study we worked on the evolution of microbial interactions in water in oil emulsions, which lead to some surprising findings that we hope to share with the scientific community soon. Keep an eye out.
Find our paper here: