||Regulation of xenobiotic efflux systems in Escherichia coli in response to environmental changes
The rising incidence of bacterial multidrug resistance has become a serious worldwide issue. The resistance-nodulation-cell division (RND)-type xenobiotic efflux system plays a major role in multidrug resistance of gram-negative bacteria. The only constitutively expressed RND system of Escherichia coli consists of the inner membrane transporter AcrB, the membrane fusion protein AcrA, and the outer membrane channel TolC. The latter two components are shared with another RND-type transporter AcrD, the expression of which is induced by environmental stimuli. In CHAPTER II of this study, I demonstrate how the RND-type ternary complexes, spanning the two membranes and the cell wall, are formed in vivo. Most foci of AcrB fused to green fluorescent protein (GFP) were fixed in the presence of TolC but showed lateral displacements when tolC was deleted. The fraction of fixed AcrB-GFP foci decreased with increasing levels of AcrD. I therefore propose that AcrD replaces AcrB in the complex with AcrA and TolC through a process I named "transporter exchange." Moreover, transporter exchange was suppressed by AcrB-specific substrates, suggesting that the ternary complex is stabilized when it is in action. These results suggest that the assembly of the RND-type efflux system is dynamically regulated in response to external stimuli, shedding new light on bacterial adaptive antibiotic resistance. Indole plays important roles in a wide variety of bacterial physiology such as virulence, drug resistance, persister cell formation, motility, and quorum sensing, among other diverse physiological processes. The expression of the inner membrane transporter gene acrD is induced by the presence of high concentrations of indole via a two-component system (TCS) consisting of the sensor kinase BaeS and the response regulator BaeR. In CHAPTER III of this study, I first observed GFP-fused AcrD encoded by the chromosomal gene with the native promoter. Observation of AcrD-GFP foci required very high concentrations of indole. In contrast, AcrD-GFP foci were readily detected in a fo/C-deletion derivative strain without exposing to exogenous indole. The deletion of the tryptophanase gene tnaA or the TCS genes baeSR abolished the constitutive expression of AcrD-GFP in the fo/C-deleted strain. I next measured indole concentrations in the presence and absence of TolC. The intracellular concentration of indole in the fo/C-deleted strain was about three times as high as that in the parental strain. In contrast, TolC had little effect on the extracellular concentration of indole. These results suggest that TolC is involved in indole efflux. Moreover, I examined whether the interaction between TolC and an inner membrane transporter is required for indole efflux. A mutant TolC protein, defective in coupling to the inner membrane transporter, decreased the level of accumulation of intracellular indole, though not as effectively as the wild-type protein. It was suggested that TolC provides a channel not only for drugs but also for indole to diffuse out of cell, and that E. coli monitors changes in the intracellular concentration of indole rather than extracellular one. Fluorescent proteins such as GFP have been used for labeling a cellular proteins. However, their large molecular weight of fluorescent proteins often interfere with the function of host protein, which would lead us to misdirected interpretations. In CHAPTER IV of this study, I applied florescent labeling methods other than fluorescent proteins to xenobiotic efflux and environment signaling systems. The fluorescein arsenical hairpin (FlAsH) specifically binds to a short peptide sequence of Cys-Cys-Pro-Gly-Cys-Cys (TC-tag), and it becomes fluorescent only when it binds covalently to a TC-tag. The TC-tagged outer membrane channel TolC of a part of the RND efflux system was labeled by FlAsH. This method resulted in the lateral localization of TolC without loss of their host functions. Moreover, I also used this FlAsH labeling method for the cytoplasmic signaling protein of Vibrio cholerae. I observed that the GFP-fused histidine kinase CheAl of V. cholerae was localized to a cell pole and lateral region of the membrane with standing incubation or in the presence of sodium azide (NaN3) that inhibits cytochrome c oxydase, whereas the protein was diffused in cellular cytosol without such treatments. The localization pattern of TC-tagged CheAl was consistent with those of the GFP fusions, arguing strongly that the behaviors of CheAl-GFP reflects the localization of the native protein. Next, I visualized the taurine/amino acid chemoreceptor Mlp37 of V. cholerae in vivo by using the fluorescently labeled L-serine 5(6)-carboxyfluorescein ester (Ser-FAM). Upon treatment with Ser-FAM, fluorescent spots were detected at poles of cells expressing wild-type mlp37. Moreover, fluorescent spots of Ser-FAM were decreased in the presence of non-fluorescent attractants serine and taurine, but not the weakest attractant L-glutamate, suggesting that the former attractants compete with Ser-FAM for binding to Mlp37. These results provide, for the first time, a tool to visualize direct ligand binding to a bacterial chemoreceptor in vivo and can also be applied to visualize signaling of TCSs and substrate transport in the xenobiotic efflux systems.