Escherichia coli population heterogeneity: Subpopulation dynamics at super-optimal temperatures

In the past years, we explored the dynamics of Escherichia coli K12 at super-optimal temperatures under static and dynamic temperature conditions ( Van Derlinden et al., 2008b, Van Derlinden et al., 2009 and Van Derlinden et al., 2010). Disturbed sigmoid growth curves, i.e., a sequence of growth, inactivation and re-growth, were observed, especially close to the maximum growth temperature. Based on the limited set of experiments (i.e., 2 static temperatures and 2 dynamic temperature profiles), the irregular growth curves were explained by postulating the co-existence of two subpopulations: a more resistant, growing population and a temperature sensitive, inactivating population.In this study, the dynamics of the two subpopulations are studied rigorously at 11 constant temperature levels in the region between 45 °C and 46 °C, with at least five repetitions per temperature.At all temperatures, the total population follows a sequence of growth, inactivation and re-growth. The sequence of different stages in the growth curves can be explained by the two subpopulations. The first growth phase and the inactivation phase reflect the presence of the sensitive subpopulation. Hereafter, the population’s dynamics are dominated by the growth of the resistant subpopulation. Generally, cell counts are characterized by a large variability.The dynamics of the two subpopulations are carefully analyzed using a heterogeneous subpopulation type model to study the relation between the kinetic parameters of the two subpopulations and temperature, and to evaluate if the fraction d of resistant cells varies with temperature. Results indicate that the growth rate of the sensitive subpopulation decreases with increasing temperature within the range of 45–46 °C. Furthermore, results point in the direction that the duration of this initial growth phase is approximately constant, i.e., around 2 h. Possibly, the stress resistance of the cells decreases after a certain period because the metabolism is fully adapted to exponential growth. Also, the growth rate of the resistant subpopulation decreases with increasing temperature. Due to the extreme variability in the cell density data, derivation of accurate relations was not possible. From the heterogeneous model implementations, given the experimental set-up, both a constant d value and a temperature dependent d value seem plausible.