A quantitative framework for bacterial competition during starvation

Bacterial communities often spend long periods under starvation, where survival depends not only on their intrinsic ability to withstand stress but also on nutrients released by dying neighbors. This creates a distinct form of competition: cells compete for recycled necromass, and the outcome should depend on physiological traits that determine nutrient uptake and maintenance demand. Here, we develop a quantitative framework for this competition using Escherichia coli populations whose starvation physiology is tuned by prior growth history. Fast-grown populations have higher maintenance demands and die slightly faster in monoculture, whereas slow-grown populations are better adapted to starvation. In co-culture, these physiological differences are strongly amplified in a frequency-dependent manner: less-adapted populations die several-fold faster than in monoculture, whereas well-adapted populations can reduce their death rate below that of stationary-phase adapted monocultures. We explain these dynamics with a shared-energy-pool model in which death releases recyclable nutrients, surviving cells consume them for maintenance, and intracellular energy sets death rate. Using independently measured parameters, the model makes parameter-free predictions for competitive survival. The predicted instantaneous death rates collapse onto a universal function of the population ratio over four orders of magnitude. Our results establish necromass recycling as a quantitative basis for bacterial competition during starvation and lay the foundation for modeling communities during starvation.