Mathematical modeling of the genetic regulatory network controlling competence for natural transformation in Streptococcus thermophilus

Bacteria are everywhere, and can be either highly beneficial or dangerous for humans. To evolve, they have developed sophisticated systems enhancing their genetic plasticity. Competence for natural transformation is a horizontal gene transfer mechanism enabling competent cells to stably acquire exogenous DNA. It is a transient physiological state, characterized in streptococci by a high level of ComX alternative sigma-factor, the master regulator of competence, which re-directs the transcription program towards the activation of genes involved in DNA acquisition. Because of the energetic cost of competence, ComX production is tightly regulated. The ComRS signaling system controls comX expression in the industrial dairy species Streptococcus thermophilus. Interestingly, the transformability of this species varies dramatically between different strains. This work aims at shedding light on the poorly understood inner workings of the ComRS system, by combining mathematical modeling and experimental approaches in order to uncover the topological hierarchy and dynamics of the timing-device function of the system, and to unravel the shut-off mechanism of competence in S. thermophilus. Moreover, the model aimed at better understanding the variability in competence levels observed among different strains of this species. For that purpose, we built a deterministic, population-scaled model of the time-course evolution of proximal players regulating comX expression. The model equations were designed to capture all the a priori knowledge available about the ComRS proximal system. First, an original forth-and-back learning step was used to calibrate the model parameters, by using a set of experimental data consisting in transcriptional profiles obtained from luciferase reporter strains in various mutant backgrounds. Second, an independent set of experimental data was used in order to assess the model quality, by testing its robustness against varying simulated scenarios, and its ability to predict the system behavior in untested conditions. The validation was performed at three levels: topological organization, kinetic properties, and reactivity of the system upon addition of the ComS signaling peptide. Once fully validated, the model was used to infer new knowledge about the ComRS system. In silico simulations and experimental data revealed that ComR is the main limiting factor for comX expression. Moreover, our analysis strongly suggests that the level of ComR production could explain the observed heterogeneity in competence levels between different strains of S. thermophilus. The model was next challenged by a subsequent in silico analysis, totally computational, focusing on the stability analysis of the steady points of the system. Two complementary approaches were proposed to determine the behavior of the system around that points. We showed that the system has two steady points: the "ON" state, corresponding to competence, and the "OFF" state, corresponding to the vegetative state. Mathematical work showed that in the case of S. thermophilus, the ComRS system ON state was unstable, while its OFF state was stable.