Formation of a Termination-Resistant RNA Polymerase Complex: Studies on the Phage 82 Q Protein

Abstract

Due to their early discovery, relative simplicity, and well-defined function, the Q antiterminator proteins of bacteriophage lambda and related lambdoid phages serve as a model for understanding how the elongation and termination behavior of RNA polymerase (RNAP) can be modified. Q initially engages E. coli RNAP that has paused immediately downstream of the phage late gene promoter PR', making contacts with both the RNAP and a DNA element embedded within PR' (the Q Binding Element, or QBE). Following successful engagement, Q remains associated with RNAP as it escapes the pause, rendering the elongation complex (EC) resistant to termination signals that otherwise prevent transcription of the phage late genes. Although Q can modify RNAP on its own, it is assisted by the essential E. coli elongation factor NusA, which facilitates Q engagement and is thought to persist as a stable component of the Q-modified EC. In this thesis, I describe a series of studies directed towards achieving a better understanding of the formation of the Q-modified EC. Focusing on the Q protein from phage 82 (82Q), I first describe an investigation of the effect of position on the ability of Q to engage the paused EC (Chapter 2). We developed an in vitro transcription assay that allowed us to compare Q engagement at the native promoter-proximal position and a novel promoter-distal position. We found that the relevance of EC positioning was unveiled by addition of the NusA protein, which stimulates Q-dependent terminator read-through when 82Q engages the promoter-proximal EC and inhibits it when 82Q engages the promoter-distal EC. Furthermore, by shortening the long nascent RNA of the promoter-distal EC, we demonstrate that nascent RNA length determines the effect of NusA on 82Q engagement. These results suggest that one function of the promoter-proximal pause is to trap the EC at a nascent RNA length that permits functional cooperation between NusA and Q. In Chapter 3, I describe a genetic approach that allowed us to rapidly identify, counterscreen, and categorize 82Q mutants. We identified two mutant classes based on the results of our analyses, one that could no longer bind the QBE and a second that was unable to modify the RNAP. An attempt to predict the 82Q structure using homology modeling produced a high-confidence structural model of 82Q (residues 82-209) based on the crystal structure of a portion of the lambdaQ protein (residues 62-206), despite relatively low sequence identity. This structural model revealed a striking structural overlap between the positions of 82Q mutations and the positions of lambdaQ mutations with analogous phenotypes, strengthening our confidence in both our 82Q mutant classification and the 82Q homology model.