Fig. 6: Growth rate under rapid nutrient fluctuations is sensitive to timescale and higher than predicted by a null model based on single shifts.

a Instantaneous growth rate dynamics modeled using growth rate responses measured from single nutrient shifts. The dynamics modeled for two slow nutrient timescales (T = 12 and 96 h) are illustrated here. At these timescales, each phase of C_high or C_low is substantially longer than the 3 or 5 h required for cells to reach either steady-state G_high or G_low after a single up- or downshift, respectively. The models used for growth rate dynamics at faster fluctuation timescales are fully described in Supplementary Fig. 10a, b and Construction of Null Model in the Supplementary information. For each modeled timescale, the growth rate dynamics were time-averaged to calculate the predicted average growth rates, G_fluc, plotted in (b). b Growth rate under rapid nutrient fluctuation (G_fluc, blue) is higher than predicted from data on single nutrient shifts between C_high and C_low. The plot shows G_fluc as a fraction of the growth rate in the steady average nutrient environment (G_ave; gold). Each measured point represents the time-averaged growth rate G and standard deviation of the mean among replicates (n = 3–4). Predicted values of G_fluc (gray) reach a maximum of G_J when the fluctuating nutrient timescale is infinitely long relative to the time required to transition from growth at G_low to G_high, and vice versa. The deviation between measured and predicted G_fluc differentiates the growth physiology at rapid fluctuation timescales from the growth behaviors expected from single shifts, and highlights the timescale-dependent nature of the growth advantage conferred by a physiology adapted for growth under rapid nutrient fluctuations.