The mechanism of oxygen sensing by the global transcriptional regulator FNR

Bacteria inhabit almost every environmental niche on Earth, including some that are so harsh that many other forms of life cannot survive. This success is at least in part due to the ability of bacteria to adapt to changes in environment, and this adaptation is rooted in their capacity to alter patterns of gene expression in response to external and internal cues. A key environmental parameter that is monitored by many bacteria is oxygen concentration. We are particularly interested in the bacterium Escherichia coli (E. coli). One of its remarkable properties is that it is able to thrive both in the presence and absence of oxygen. To do this it has to dramatically alter its metabolism, but this has consequences because without oxygen the potential for energy conservation and growth are limited compared to when oxygen is present. To test whether oxygen is present E. coli uses a protein called FNR, which acts as an oxygen sensor. It has a special co-factor called an iron-sulfur cluster that reacts with oxygen in a way that switches FNR off. In the 'off-state' (no iron-sulfur cluster) FNR cannot bind to DNA to activate expression of genes that are used during growth in the absence of oxygen. When there is no oxygen the iron-sulfur remains associated with FNR and so the protein can bind to DNA and activate expression of genes that are needed for growth in the absence of oxygen. For the last few years we have studied the reaction of oxygen with the FNR iron-sulfur cluster. These studies have revealed the complex biochemistry of the reaction and how this makes the FNR protein an exquisitely sensitive oxygen sensor. However, our studies have raised even more questions. We now propose to address some of these by determining the molecular choreography of the reaction of oxygen with the FNR iron-sulfur cluster and how this leads to altered DNA-binding properties. We will also investigate how the potentially toxic reaction products are managed by the cell, and why some bacteria have more than one FNR protein; could it be that different FNR proteins have evolved to operate in different ranges of oxygen concentration? The work is important at several levels. It will allow us to understand more about a basic biological process that is fundamental for the virulent properties of many bacterial pathogens, such as Salmonella and E. coli. It will provide new insight and a deeper understanding of an ancient class of proteins that evolved in an anaerobic world. It has wider significance for how signal perception and signal transduction are linked in a ubiquitous family of bacterial gene regulators. And finally, it offers possibilities of designing molecular switches that will respond to changes in oxygen levels for use in biotechnology.

The bacterial transcriptional regulator FNR, which orchestrates the switch between aerobic and anaerobic respiration, consists of two distinct domains that provide DNA-binding (C-terminal) and sensory (N-terminal) functions. FNR becomes active under anaerobic conditions through the binding of a [4Fe-4S] cluster to the sensory domain, leading to dimerization and DNA-binding. O2 triggers the conversion of the [4Fe-4S] cluster to a [2Fe-2S] form, causing the protein to monomerise with the loss of specific DNA-binding. Recently, we showed that cluster conversion proceeds in two distinct steps, resulting in a [3Fe-4S] intermediate, and that the kinetics of the reaction are modulated by the protein environment. This work has raised a raft of new questions that now need to be addressed. Using electrochemical methods, we will investigate the reaction between the [4Fe-4S]2+ cluster and O2 that initiates cluster conversion and leads to the formation of the [3Fe-4S] intermediate, and with Resonance Raman spectroscopy we will build on highly significant preliminary data indicating that released cluster sulfide may be stored within the protein. We will investigate to what extent the [3Fe-4S] and [2Fe-2S] cluster forms can undergo a reverse reaction leading to repair of the [4Fe-4S] cluster. We will also explore the mechanism of signal transduction from the sensory domain to the DNA-binding domain to address key issues such as the effect of DNA binding on FNR oxygen sensitivity. Some bacteria contain multiple FNR proteins and we have proposed that FNR proteins may have evolved to respond to different oxygen concentrations. We will test this by studying in detail three FNR proteins from the same bacterium. Finally, we have generated a systems-level steady-state model of the dynamics of FNR in the cell, which we propose to test and develop further. Together, this work will provide an unprecedented level of understanding of FNR at a molecular, systems and evolutionary level.