A genetic dissection of the interactions between the CbtA toxin of Escherichia coli and the bacterial cytoskeleton

Abstract

Prokaryotic chromosomal toxin-antitoxin (TA) systems, consisting of a stable toxin and a labile, cotranscribed antitoxin, have been shown to target a number of essential processes in bacteria. The Escherichia coli genome encodes multiple chromosomal TA systems, including a family of three homologous systems, cbtA/cbeA, ykfI/yafW, and ypjF/yfjZ, that targets the bacterial cytoskeleton. Upon overproduction of the CbtA, YkfI, or YpjF toxins, E. coli cells adopt a lemon-like morphology, reminiscent of that seen upon simultaneous inhibition of cell division and cell elongation pathways. Consistent with this observed morphology, previously published work has shown that the CbtA toxin interacts with both the tubulin-homolog FtsZ, the master regulator of cell division, and the actin-homolog MreB, an essential cell elongation factor. Despite these findings, it has not been demonstrated that the interactions of CbtA with MreB and FtsZ are directly responsible for its observed cellular toxicity, nor have any mechanistic details been revealed. The goal of this research was to elucidate the molecular basis for CbtA toxicity by genetically characterizing the interactions between CbtA and its cytoskeletal targets.

By means of a transcription-based bacterial two-hybrid system, we can detect interactions between CbtA and both FtsZ and MreB. In Chapter 2 of this dissertation, I describe the isolation and characterization of two CbtA mutants, each of which interacts with only a single cytoskeletal target. CbtA-F65S is unable to bind FtsZ, but maintains interaction with MreB, whereas CbtA-R15C exhibits the reverse interaction profile. Morphological observation of cells overproducing each of these variants establishes that the observed effects of CbtA on cell division and cell elongation are genetically separable. I show further that in combination, the substitutions F65S and R15C alleviate the toxicity of CbtA, consistent with the premise that CbtA toxicity depends on its interactions with both FtsZ and MreB.

In Chapters 3 and 4, I describe our efforts to map the CbtA-binding determinants on FtsZ and MreB, respectively. Through bacterial two-hybrid studies and the development of a Bacillus subtilis heterologous system, we have identified the H6/H7 loop of FtsZ as the CbtA- binding site and provided additional evidence that the CbtA-FtsZ interaction contributes directly to the observed cellular toxicity. Two-hybrid studies indicate that YkfI and YpjF also interact with the H6/H7 loop of FtsZ, and we believe that this represents a new FtsZ inhibitory interaction. Based on FtsZ structural studies, the H6/H7 loop is thought to be important for FtsZ protofilament formation, and thus we propose that CbtA binding to this surface loop blocks FtsZ polymerization. Furthermore, residues in this loop have been implicated in FtsZ bundling, suggesting that CbtA may also inhibit the formation of stabilizing lateral interactions. We employed similar genetic methods to characterize the interaction between MreB and CbtA. Two-hybrid and morphology data presented in Chapter 4 of this dissertation suggest that CbtA interacts with the flat MreB surface that mediates formation of the functionally active MreB double filament. In total, the genetic analysis presented in this dissertation establishes that CbtA toxicity depends on independent interactions of CbtA with the tubulin homolog FtsZ and the actin homolog MreB, as well as providing molecular insight into the basis of this dual inhibitory action.