Improved growth of Escherichia coli in aminoglycoside antibiotics by the zor-orz toxin-antitoxin system.

Type I toxin-antitoxin systems consist of a small protein (under 60 amino acids) whose overproduction can result in cell growth stasis or death, and a small RNA that represses translation of the toxin mRNA. Despite their potential toxicity, type I toxin proteins are increasingly linked to improved survival of bacteria in stressful environments and antibiotic persistence. While the interaction of toxin mRNAs with their cognate antitoxin sRNAs in some systems are well characterized, additional translational control of many toxins and their biological roles are not well understood. Using an ectopic overexpression system, we show that the efficient translation of a chromosomally encoded type I toxin, ZorO, requires mRNA processing of its long 5' untranslated region (UTR; Δ28 UTR). The severity of ZorO induced toxicity on growth inhibition, membrane depolarization, and ATP depletion were significantly increased if expressed from the Δ28 UTR versus the full-length UTR. ZorO did not form large pores as evident via a liposomal leakage assay, in vivo morphological analyses, and measurement of ATP loss. Further, increasing the copy number of the entire zor-orz locus significantly improved growth of bacterial cells in the presence of kanamycin and increased the minimum inhibitory concentration against kanamycin and gentamycin; however, no such benefit was observed against other antibiotics. This supports a role for the zor-orz locus as a protective measure against specific stress agents and is likely not part of a general stress response mechanism. Combined, these data shed more insights into the possible native functions for type I toxin proteins. IMPORTANCE Bacterial species can harbor gene pairs known as type I toxin-antitoxin systems where one gene encodes a small protein that is toxic to the bacteria producing it and a second gene that encodes a small RNA antitoxin to prevent toxicity. While artificial overproduction of type I toxin proteins can lead to cell growth inhibition and cell lysis, the endogenous translation of type I toxins appears to be tightly regulated. Here, we show translational regulation controls production of the ZorO type I toxin and prevents subsequent negative effects on the cell. Further, we demonstrate a role for zorO and its cognate antitoxin in improved growth of E. coli in the presence of aminoglycoside antibiotics.

Translocation of gut commensal bacteria to the brain.

The gut-brain axis, a bidirectional signaling network between the intestine and the central nervous system, is crucial to the regulation of host physiology and inflammation. Recent advances suggest a strong correlation between gut dysbiosis and neurological diseases, however, relatively little is known about how gut bacteria impact the brain. Here, we reveal that gut commensal bacteria can translocate directly to the brain when mice are fed an altered diet that causes dysbiosis and intestinal permeability, and that this also occurs without diet alteration in distinct murine models of neurological disease. The bacteria were not found in other systemic sites or the blood, but were detected in the vagus nerve. Unilateral cervical vagotomy significantly reduced the number of bacteria in the brain, implicating the vagus nerve as a conduit for translocation. The presence of bacteria in the brain correlated with microglial activation, a marker of neuroinflammation, and with neural protein aggregation, a hallmark of several neurodegenerative diseases. In at least one model, the presence of bacteria in the brain was reversible as a switch from high-fat to standard diet resulted in amelioration of intestinal permeability, led to a gradual loss of detectable bacteria in the brain, and reduced the number of neural protein aggregates. Further, in murine models of Alzheimer's disease, Parkinson's disease, and autism spectrum disorder, we observed gut dysbiosis, gut leakiness, bacterial translocation to the brain, and microglial activation. These data reveal a commensal bacterial translocation axis to the brain in models of diverse neurological diseases.

Charged Amino Acids Contribute to ZorO Toxicity.

Chromosomally encoded toxin-antitoxin systems have been increasingly identified and characterized across bacterial species over the past two decades. Overproduction of the toxin gene results in cell growth stasis or death for the producing cell, but co-expression of its antitoxin can repress the toxic effects. For the subcategory of type I toxin-antitoxin systems, many of the described toxin genes encode a small, hydrophobic protein with several charged residues distributed across the sequence of the toxic protein. Though these charged residues are hypothesized to be critical for the toxic effects of the protein, they have not been studied broadly across different type I toxins. Herein, we mutated codons encoding charged residues in the type I toxin zorO, from the zor-orz toxin-antitoxin system, to determine their impacts on growth inhibition, membrane depolarization, ATP depletion, and the localization of this small protein. The non-toxic variants of ZorO accumulated both in the membrane and cytoplasm, indicating that membrane localization alone is not sufficient for its toxicity. While mutation of a charged residue could result in altered toxicity, this was dependent not only on the position of the amino acid within the protein but also on the residue to which it was converted, suggesting a complex role of charged residues in ZorO-mediated toxicity. A previous study indicated that additional copies of the zor-orz system improved growth in aminoglycosides: within, we note that this improved growth is independent of ZorO toxicity. By increasing the copy number of the zorO gene fused with a FLAG-tag, we were able to detect the protein expressed from its native promoter elements: an important step for future studies of toxin expression and function.