Bimodality and local signaling in the c-di-GMP network of E. coli

Biofilms protect bacteria from environmental threats, such as predators, antibiotics, or attacks by host immune systems. The second messenger c-di-GMP is a key regulator of bacterial behavior and biofilm formation, orchestrating the transition from a solitary and motile to a sessile community lifestyle. The cellular concentration of c-di-GMP is regulated by diguanylate cyclases (DGCs) that produce and phosphodiesterases (PDEs) that degrade c-di-GMP in response to intra- and extracellular stimuli. Many bacteria have evolved multiple DGCs and PDEs to broaden their sensing capabilities and to respond to a large variety of signals. However, it remains poorly understood how multiple signaling pathways converge into specific and robust cellular readouts and how the hypersensitive c-di-GMP network absorbs noise emerging from stochastic expression or activation of DGCs and PDEs.
Here, I use Escherichia coli as a model to study how gradual changes and fluctuation of c-di-GMP are converted into deterministic cellular responses. In the first chapter of my thesis, I describe a novel molecular mechanism through which bacteria convert gradual changes of c-di-GMP into binary signaling outcomes. Together with my collaborators, I present genetic, biochemical, and structural evidence to demonstrate that the E. coli PDE PdeL operates as a hypersensitive switch to quench noise and to install robust bimodal c-di-GMP regimes. We show that PdeL is both a c-di-GMP degrading catalyst and a transcription factor that stimulates its own expression. In addition, PdeL is a c-di-GMP sensor as both transcriptional and enzymatic activities are high at low c-di-GMP but curbed at high intracellular c-di-GMP concentrations. PdeL adopts an inert dimer conformation at high c-di-GMP levels but switches into a highly active tetramer conformation when c-di-GMP is lowered. We show that with its highly cooperative behavior, PdeL converts populations experiencing gradual changes of c-di-GMP into bimodal populations where individual cells exhibit either high or low c-di-GMP. Based on the observation that pdeL expression is strongly hysteretic, we propose that this switch provides E. coli with a short-term memory which entails robustness to costly lifestyle transitions. Finally, I observed that PdeL effectively protects E. coli against specific bacteriophage predators, indicating that this simple molecular switch also serves as a bet-hedging device to minimize risks associated with biofilm formation.
In the second chapter of my thesis, I describe a novel regulatory pathway responsible for the synthesis and secretion of an as yet unknown extracellular glycan polymer in E. coli. Starting from the observation that bacteriophage N4 can only infect E. coli if PdeL is in its off state and c-di-GMP levels high, this part investigates how c-di-GMP contributes to phage entry.
I could demonstrate that N4 infection requires two DGCs, DgcQ and DgcJ, which sense arginine and an as yet unidentified component of complex media, respectively. Genetic data suggest that DgcJ and DgcQ expedite N4 infection by stimulating the synthesis of a novel surface-associated glycan polymer, which is used by N4 as a primary surface receptor. Genetic data combined with homology modeling identified NfrB and NfrA as inner and outer membrane components of the N4-specific polysaccharide secretion system. NfrB not only shows strong homologies to glycosyltransferases but also harbors a C-terminal MshE-like c-di-GMP binding domain. Based on this, and based on the observation that the UDP-N-acetylglucosamine 2-epimerase WecB is also essential for N4 infection, we propose that c-di-GMP activates NfrB to polymerize a glycan polymer containing N-acetylmannosamine (ManNAc), which is then secreted through the NfrA outer membrane porin to the cell surface where it serves as the primary receptor for bacteriophage N4. Preliminary data indicate that DgcJ specifically activates the NfrBA pathway by acting as a «local pacemaker» while DgcQ acts globally and stimulates this pathway by functionally interacting with DgcJ activity.
Overall, this work uncovers novel mechanistic principles, which bacteria use to convert changes of a small diffusible signaling molecules into deterministic, precise, and irreversible cellular responses. The finding that bacteria can use spatially localized signaling domains to stimulate specific cellular processes and that they are able to convert graded into binary frequency-based responses greatly expands our knowledge on the extensive signaling repertoire that bacteria have evolved to maximize their fitness in constantly changing environments. While the nature and the function of the novel glycan are yet to be understood, it is evident that E. coli has evolved the regulatory interface allowing for precise utilization of the glycan to mitigate the risk of phage invasion and other adverse effects.