This dissertation addresses the mechanism of σ54 activation by the AAA+ ATPase transcriptional activators. The first chapter provides a general introduction to σ54‑mediated bacterial transcription initiation by outlining the existing structural and biochemical data on σ54 and its activators. The majority of our structural information comes from high resolution structures of individual σ54 and transcriptional activator domains, with the rest coming from low resolution structures that determine relative positions of the domains. The structural and biochemical properties of these domains, viewed in the context of the mechanisms of related motor proteins, led to our hypothesis of σ54 activation. In this dissertation, I propose that the transcriptional activator’s hexameric ATPase domain uses conserved loops, known to contact σ54, to pull on and thread the σ54 N‑terminal activator interacting domain through its central pore. In this model, through processive rounds of ATP hydrolysis, the transcriptional activator applies enough force to generate a conformational change in the σ54‑RNA polymerase holoenzyme that allows it to melt DNA thus initiating transcription. The primary goal of my work has been the study of the activation mechanism using a variety of techniques, including traditional NMR‑based structure determination, in vivo and in vitro biochemistry, and single molecule optical tweezers experiments.
The second chapter outlines the central work of this dissertation, the structural characterization of the region of σ54 responsible for contacting the activator and necessary for initiating transcription. We show that the activator interacting domain (AID) is intrinsically disordered, but becomes ordered when bound to the transcriptional activator, NtrC1, in its ATP state. In particular, we show that two predicted helices in the sequence are sufficient for activator binding with native‑like affinity, and that the first helix in particular represents the primary region of contact. We applied TROSY‑based NMR techniques to the structure determination of this high molecular weight complex, but most signals were broad due to dynamics or disorder, preventing a high resolution structure determination of the AID bound to the activator. This indicates that the AID does not bind the activator in a single conformation, but rather in multiple conformations each experiencing a different chemical environment. This may reflect an activation mechanism involving dynamic changes to the σ54‑activator interaction site.
The third chapter takes an in vivo biochemical approach to study the σ54 activation mechanism by characterizing the functionality of σ54 with insertions and deletions near the activator interacting domain. These insertions distinguish between two competing mechanisms: whether the σ54 AID only binds the surface of the activator or, as we propose, is threaded through its central pore. We find that a flexible linker region between the sites of activator and RNA polymerase binding is essential for fully functional σ54. While the length of the linker does not matter, the insertion of a small, stably folded domain between these two interaction sites is detrimental to σ54 activity in vivo. This is consistent with the threading of the AID and linker by the transcriptional activator, which would not be inhibited by the addition of extra unstructured residues but would be stalled by the addition of a folded domain in the linker.
The fourth chapter outlines single molecule optical tweezers experiments to characterize the effect of force on a domain of σ54. In particular, we examine the possibility that the σ54 core binding domain (CBD), responsible for contacting core RNA polymerase, acts as a conformational fracture point that undergoes rearrangements when force is applied. The NMR structure of the CBD revealed two folded subdomains with a hydrophobic interface that could plausibly be disrupted by the force of the activator threading the N‑terminus. By applying force with optical tweezers to either end of the CBD alone, we observed a separate unfolding event likely corresponding to unfolding of its seventh helix, but there was no evidence that the less stable of the two CBD subdomains unfolds before the other. Future experiments to conclusively test the force‑dependent activation of σ54 may require reconstituting and pulling on the full RNA polymerase holoenzyme complex, including core RNAP, promoter DNA, and full length σ54.
The fifth chapter steps back from the study of the σ54 activation by the transcriptional activators to examine the regulation of the transcriptional activators themselves. Existing evidence shows that phosphorylation of the regulatory receiver domain of the transcriptional activator NtrC changes the population of its active and inactive states, with most states resembling the active conformation after phosphorylation. The activated receiver domain promotes the oligomerization of the ATPase domain, which in turn can activate σ54. In this chapter, I studied an NtrC receiver domain from the piezophilic bacteria, Shewanella violacea, which turns on σ54‑dependent gene expression in response to high pressure. We hypothesized that pressure alone might drive the conformational change normally associated with phosphorylation, thereby activating S.v. NtrC independent of its normal, two‑component signaling pathway. Using high pressure NMR spectroscopy, we showed that increased pressure alone does not increase the population of the S.v. NtrC receiver domain’s active state. Future work must consider the pressure sensing behavior of the other components in the NtrC signaling pathway