High-dimensional characterization of phototroph-heterotroph interactions

Interactions between phototrophic and heterotrophic microbes lie at the heart of global biogeochemistry, ecosystem productivity, and biofuel generation. These microbes are known to interact with each other via resource exchange and competition - Phototrophs provide organic carbon to heterotrophs and heterotrophs provide essential micronutrients to phototrophs while they compete with each other for inorganic nutrients such as phosphorus and nitrogen. However, we lack a quantitative understanding of how the interactions between phototrophs and heterotrophs are impacted by the chemical nature of the environment. To address this, we conducted experimental studies to assay interactions between the phototroph, Chlamydomonas reinhardtii, and the heterotroph, Escherichia coli, as a function of five environmental parameters - pH, buffering capacity,availability of three non-substitutable resources - carbon, nitrogen, and light. To aid this high-dimensional study, a high throughput droplet-based microfluidic platform that allows us to rapidly construct thousands of environmental conditions was implemented. Using this platform, we have screened for interactions in our model phototroph-heterotroph system in ~840 environmental conditions, by setting up ~200,000 microcosms on the microfluidic platform. Using machine learning, we show that the parameters pH and buffering capacity control the interaction structure in phototroph-heterotroph communities. Further, we find that the nature of the carbon source driving heterotroph growth - glycolytic or gluconeogenic alters the interaction structure via impacting pH. These results are in contrast to the prevailing view that phototroph-heterotroph interactions are governed primarily by the exchange of and competition for nutrients. Our work presents a new view of how the chemical environment impacts phototroph-heterotroph interactions.