Metabolic feedbacks drive population dynamics and can lead to oscillations among leaf bacteria

Metabolic interactions are fundamental to the assembly and function of microbiomes. Yet, our understanding of how specific interaction mechanisms can drive broader ecological outcomes and population dynamics remains limited. Here, we monitor interactions resulting from plant oligosaccharide degradation by leaf-associated bacteria using a microfluidic device that enables direct cell observation and quantitative metabolite detection. This approach enables the identification of key metabolic mediators, revealing recipient-specific patterns of carbon substrate and cofactor complementation. By linking these patterns to emergent dynamics observed between pairs of bacteria, we identify metabolically driven feedbacks that could lead to a variety of ecological outcomes - from outcompetition to coexistence characterized by oscillating population abundances. Investigating these observations with metabolic modeling allows us to systematically assess the impact of specific molecular mediators on population dynamics, yielding predictions of interaction outcomes that we validate experimentally. Our results provide a detailed mapping of metabolic mechanisms to emergent population trajectories among environmental microbes and help inform strategies for designing microbiomes with desired steady states.