Funded Grants

Mapping the microbial survival toolbox: Using dynamic age distributions to infer the behavior of individuals within populations

The adaptations of microorganisms to fluctuating environmental conditions are remarkably diverse and sophisticated. Some of the best-studied examples include the chemotaxis genetic network, which bacteria use to sense and move up nutrient gradients, and the GAL regulon by which yeast adapt to changes in sugar availability [1, 2]. Such genetic networks can be classified as a type of survival strategy, known as responsive switching, which involves sensing an environmental fluctuation and coordinating the induction and repression of specific genes in response to the change. To date, hundreds of such sensor-response modalities have been elucidated [3]. Microorganisms, however, possess another important type of survival strategy, known as stochastic switching [4, 5], which operates without any sensory system, and allows cells spontaneously to induce particular genes, and repress others. One prime example is the persistence stochastic switch in Escherichia coli, whereby cells switch spontaneously into a slow-growing state, known as the persistent phenotype, which allows survival of antibiotics and other environmental stresses [6-8]. The rate with which cells make this switch is extremely low, on the order of one switch per 105 cells per cell division. Thus, a tiny subpopulation of antibiotic-tolerant E.coli cells is maintained within a large population of antibiotic-sensitive cells. Such switches are prevalent in bacteria [9, 10] and in fungi [11-13] including in pathogenic strains and organisms, where they are believed to increase microbial survival rates under antimicrobial stresses.

We previously introduced and analyzed a population dynamics model of phenotypic switching (both stochastic and responsive) [14]. The stochastic switching population adapts to fluctuating conditions by selective expansions of small subpopulations that are already pre-adapted to the new condition, and are maintained in the population before the change occurs (see Figure 1 of Proposal). The responsive switching population does not maintain diversity, and does not rely on selective expansions for population-level adaptation. Instead, each organism independently switches to the appropriate phenotype via sensory pathways. For stochastic switches, switching rates are typically much smaller than the growth rates (i.e. phenotypic states are heritable but reversible), and exhibit little or no dependence on the environmental state; while for responsive switches, they are typically faster than growth rates and exhibit a strong dependence on the environmental state.

Both responsive and stochastic switching strategies allow cells to change their state, or phenotype, to one that is appropriate for specific environmental conditions. This leads to the following fundamental biological question: How do evolution and ecology influence the repertoire of phenotypic switching strategies employed by any given microorganism? More specifically, why are some genes regulated by stochastic switching and others by responsive switching? And, more broadly, what can the repertoire of phenotypic switching strategies employed by a microorganism tell us about its natural environment?

By developing and analyzing mathematical models, in combination with experiments, we have begun to answer these complex questions. For microorganisms, there exists a strong evolutionary constraint on the maintenance of sensory networks