Rapid Curtailing of the Stringent Response by Toxin-Antitoxin Module-Encoded mRNases

Existence of log-phase Escherichia coli persisters and lasting memory of a starvation pulse

The authors characterize the growth condition dependence of survival of bacteria exposed to lethal antibiotics for a week. 1-h starvation pulse is shown to cause an increase in survival for days.

The vast majority of a bacterial population is killed when treated with a lethal concentration of antibiotics. The time scale of this killing is often comparable with the bacterial generation time before the addition of antibiotics. Yet, a small subpopulation typically survives for an extended period. However, the long-term killing dynamics of bacterial cells has not been fully quantified even in well-controlled laboratory conditions. We constructed a week-long killing assay and followed the survival fraction of Escherichia coli K12 exposed to a high concentration of ciprofloxacin. We found that long-term survivors were formed during exponential growth, with some cells surviving at least 7 d. The long-term dynamics contained at least three time scales, which greatly enhances predictions of the population survival time compared with the biphasic extrapolation from the short-term behavior. Furthermore, we observed a long memory effect of a brief starvation pulse, which was dependent on the (p)ppGpp synthase relA. Specifically, 1 h of carbon starvation before antibiotics exposure increased the surviving fraction by nearly 100-fold even after 4 d of ciprofloxacin treatment.

Valine-Induced Isoleucine Starvation in Escherichia coli K-12 Studied by Spike-In Normalized RNA Sequencing

Escherichia coli cells respond to a period of famine by globally reorganizing their gene expression. The changes are known as the stringent response, which is orchestrated by the alarmone ppGpp that binds directly to RNA polymerase. The resulting changes in gene expression are particularly well studied in the case of amino acid starvation. We used deep RNA sequencing in combination with spike-in cells to measure global changes in the transcriptome after valine-induced isoleucine starvation of a standard E. coli K12 strain. Owing to the whole-cell spike-in method that eliminates variations in RNA extraction efficiency between samples, we show that ribosomal RNA levels are reduced during isoleucine starvation and we quantify how the change in cellular RNA content affects estimates of gene regulation. Specifically, we show that standard data normalization relying on sample sequencing depth underestimates the number of down-regulated genes in the stringent response and overestimates the number of up-regulated genes by approximately 40%. The whole-cell spike-in method also made it possible to quantify how rapidly the pool of total messenger RNA (mRNA) decreases upon amino acid starvation. A principal component analysis showed that the first two components together described 69% of the variability of the data, underlining that large and highly coordinated regulons are at play in the stringent response. The induction of starvation by sudden addition of high valine concentrations provoked prominent regulatory responses outside of the expected ppGpp, RpoS, and Lrp regulons. This underlines the notion that with the high resolution possible in deep RNA sequencing analysis, any different starvation method (e.g., nitrogen-deprivation, removal of an amino acid from an auxotroph strain, or valine addition to E. coli K12 strains) will produce measurable variations in the stress response produced by the cells to cope with the specific treatment.

Cysteine homeostasis under inhibition of protein synthesis in Escherichia coli cells

Increased intracellular cysteine poses a potential danger to cells due to the high ability of cysteine to reduce free iron and promote the Fenton reaction. Here, we studied ways to maintain cysteine homeostasis in E. coli cells while inhibiting protein synthesis with valine or chloramphenicol. When growing wild-type bacteria on minimal medium with sulfate, an excess of cysteine resulting from the inhibition of protein synthesis is mainly incorporated into glutathione (up to 90%), which, therefore, can be considered as cysteine buffer. The share of hydrogen sulfide, which is the product of cysteine degradation by cysteine synthase B (CysM), does not exceed 1-3%, the rest falls on free cysteine, exported from cells. As a result, intracellular free cysteine is maintained at a low level (about 0.1 mM). The lack of glutathione in the gshA mutant increases H₂S production and excretion of cysteine and leads to a threefold increase in the level of intracellular cysteine in response to valine and chloramphenicol. The relA mutants, exposed to valine, produce more H₂S, dramatically accelerate the export of glutathione and accumulate more cysteine in the cytoplasm than their parent, which indicates that the regulatory nucleotide (p)ppGpp is involved in maintaining cysteine homeostasis. Disruption of cysteine homeostasis in gshA and relA mutants increases their sensitivity to peroxide stress.

Intramolecular Interactions Dominate the Autoregulation of Escherichia coli Stringent Factor RelA

Amino acid starvation in Escherichia coli activates the enzymatic activity of the stringent factor RelA, leading to accumulation of the alarmone nucleotide (p)ppGpp. The alarmone acts as an intercellular messenger to regulate transcription, translation and metabolism to mediate bacterial stress adaptation. The enzymatic activity of RelA is subject to multi-layered allosteric control executed both by ligands - such as "starved" ribosomal complexes, deacylated tRNA and pppGpp - and by individual RelA domains. The auto-regulation of RelA is proposed to act either in cis (inhibition of the enzymatic activity of the N-terminal region, NTD, by regulatory C-terminal region, CTD) or in trans (CTD-mediated dimerization leading to enzyme inhibition). In this report, we probed the regulatory roles of the individual domains of E. coli RelA and our results are not indicative of RelA dimerization being the key regulatory mechanism. First, at growth-permitting levels, ectopic expression of RelA CTD does not interfere with activation of native RelA, indicating lack of regulation via inhibitory complex formation in the cell. Second, in our biochemical assays, increasing RelA concentration does not decrease the enzyme activity, as would be expected in the case of efficient auto-inhibition via dimerization. Third, while high-level CTD expression efficiently inhibits the growth, the effect is independent of native RelA and is mediated by direct inhibition of protein synthesis, likely via direct interaction with the ribosomal A-site. Finally, deletion of the RRM domain of the CTD region leads to growth inhibition mediated by accumulation of (p)ppGpp, suggesting de-regulation of the synthetic activity in this mutant.

Serine-Threonine Kinases Encoded by Split hipA Homologs Inhibit Tryptophanyl-tRNA Synthetase

Bacterial toxin-antitoxin (TA) modules confer multidrug tolerance (persistence) that may contribute to the recalcitrance of chronic and recurrent infections. The first high-persister gene identified was hipA of Escherichia coli strain K-12, which encodes a kinase that inhibits glutamyl-tRNA synthetase. The hipA gene encodes the toxin of the hipBA TA module, while hipB encodes an antitoxin that counteracts HipA. Here, we describe a novel, widespread TA gene family, hipBST, that encodes HipT, which exhibits sequence similarity with the C terminus of HipA. HipT is a kinase that phosphorylates tryptophanyl-tRNA synthetase and thereby inhibits translation and induces the stringent response. Thus, this new TA gene family may contribute to the survival and spread of bacterial pathogens.

Type II toxin-antitoxin (TA) modules encode a stable toxin that inhibits cell growth and an unstable protein antitoxin that neutralizes the toxin by direct protein-protein contact. hipBA of Escherichia coli strain K-12 codes for HipA, a serine-threonine kinase that phosphorylates and inhibits glutamyl-tRNA synthetase. Induction of hipA inhibits charging of glutamyl-tRNA that, in turn, inhibits translation and induces RelA-dependent (p)ppGpp synthesis and multidrug tolerance. Here, we describe the discovery of a three-component TA gene family that encodes toxin HipT, which exhibits sequence similarity with the C-terminal part of HipA. A genetic screening revealed that trpS in high copy numbers suppresses HipT-mediated growth inhibition. We show that HipT of E. coli O127 is a kinase that phosphorylates tryptophanyl-tRNA synthetase in vitro at a conserved serine residue. Consistently, induction of hipT inhibits cell growth and stimulates production of (p)ppGpp. The gene immediately upstream from hipT, called hipS, encodes a small protein that exhibits sequence similarity with the N terminus of HipA. HipT kinase was neutralized by cognate HipS in vivo, whereas the third component, HipB, encoded by the first gene of the operon, did not counteract HipT kinase activity. However, HipB augmented the ability of HipS to neutralize HipT. Analysis of two additional hipBST-homologous modules showed that, indeed, HipS functions as an antitoxin in these cases also. Thus, hipBST constitutes a novel family of tricomponent TA modules where hipA has been split into two genes, hipS and hipT, that function as a novel type of TA pair.

(p)ppGpp-mediated stress response induced by defects in outer membrane biogenesis and ATP production promotes survival in Escherichia coli

Cellular growth requires a high level of coordination to ensure that all processes run in concert. The role of the nucleotide alarmone (p)ppGpp has been extensively studied in response to external stresses, such as amino acid starvation, in Escherichia coli, but much less is known about the involvement of (p)ppGpp in response to perturbations in intracellular processes. We therefore employed CRISPRi to transcriptionally repress essential genes involved in 14 vital processes and investigated whether a (p)ppGpp-mediated response would be induced. We show that (p)ppGpp is produced and required for a pertinent stress response during interference with outer membrane biogenesis and ADP synthesis specifically. When these processes were perturbed via the transcriptional repression of essential genes, wild type E. coli MG1655 ceased growing and entered a semi-dormant state, whereas isogenic (p)ppGpp⁰ cells continued to grow uncontrollably to the point of lysis. Furthermore, in vivo measurements revealed that the ATP levels were intrinsically offset in (p)ppGpp⁰ cells, further indicating a role for the alarmone in cellular energy homeostasis. In summary, our investigation suggests that (p)ppGpp acts as a coordinator of cell growth in response to imbalances in outer membrane biogenesis and adenosine ribonucleotide synthesis, elucidating novel roles for (p)ppGpp in bacterial physiology.

Activation of the Stringent Response by Loading of RelA-tRNA Complexes at the Ribosomal A-Site

RelA/SpoT homologs (RSHs) are ubiquitous bacterial enzymes that synthesize and hydrolyze (p)ppGpp in response to environmental challenges. Bacteria cannot survive in hosts and produce infection without activating the (p)ppGpp-mediated stringent response, but it is not yet understood how the enzymatic activities of RSHs are controlled. Using UV crosslinking and deep sequencing, we show that Escherichia coli RelA ((p)ppGpp synthetase I) interacts with uncharged tRNA without being activated. Amino acid starvation leads to loading of cognate tRNA⋅RelA complexes at vacant ribosomal A-sites. In turn, RelA is activated and synthesizes (p)ppGpp. Mutation of a single, conserved residue in RelA simultaneously prevents tRNA binding, ribosome binding, and activation of RelA, showing that all three processes are interdependent. Our results support a model in which (p)ppGpp synthesis occurs by ribosome-bound RelA interacting with the Sarcin-Ricin loop of 23S rRNA.

Transfer RNA instability as a stress response in Escherichia coli: Rapid dynamics of the tRNA pool as a function of demand

Production of the translation apparatus of E. coli is carefully matched to the demand for protein synthesis posed by a given growth condition. For example, the fraction of RNA polymerases that transcribe rRNA and tRNA drops from 80% during rapid growth to 24% within minutes of a sudden amino acid starvation. We recently reported in Nucleic Acids Research that the tRNA pool is more dynamically regulated than previously thought. In addition to the regulation at the level of synthesis, we found that tRNAs are subject to demand-based regulation at the level of their degradation. In this point-of-view article we address the question of why this phenomenon has not previously been described. We also present data that expands on the mechanism of tRNA degradation, and we discuss the possible implications of tRNA instability for the ability of E. coli to cope with stresses that affect the translation process.

Novel (p)ppGpp Binding and Metabolizing Proteins of Escherichia coli

The alarmone (p)ppGpp plays pivotal roles in basic bacterial stress responses by increasing tolerance of various nutritional limitations and chemical insults, including antibiotics. Despite intensive studies since (p)ppGpp was discovered over 4 decades ago, (p)ppGpp binding proteins have not been systematically identified in Escherichia coli. We applied DRaCALA (differential radial capillary action of ligand assay) to identify (p)ppGpp-protein interactions. We discovered 12 new (p)ppGpp targets in E. coli that, based on their physiological functions, could be classified into four major groups, involved in (i) purine nucleotide homeostasis (YgdH), (ii) ribosome biogenesis and translation (RsgA, Era, HflX, and LepA), (iii) maturation of dehydrogenases (HypB), and (iv) metabolism of (p)ppGpp (MutT, NudG, TrmE, NadR, PhoA, and UshA). We present a comprehensive and comparative biochemical and physiological characterization of these novel (p)ppGpp targets together with a comparative analysis of relevant, known (p)ppGpp binding proteins. Via this, primary targets of (p)ppGpp in E. coli are identified. The GTP salvage biosynthesis pathway and ribosome biogenesis and translation are confirmed as targets of (p)ppGpp that are highly conserved between E. coli and Firmicutes. In addition, an alternative (p)ppGpp degradative pathway, involving NudG and MutT, was uncovered. This report thus significantly expands the known cohort of (p)ppGpp targets in E. coli.

Antibiotic resistance and tolerance exhibited by pathogenic bacteria have resulted in a global public health crisis. Remarkably, almost all bacterial pathogens require the alarmone (p)ppGpp to be virulent. Thus, (p)ppGpp not only induces tolerance of nutritional limitations and chemical insults, including antibiotics, but is also often required for induction of virulence genes. However, understanding of the molecular targets of (p)ppGpp and the mechanisms by which (p)ppGpp influences bacterial physiology is incomplete. In this study, a systematic approach was used to uncover novel targets of (p)ppGpp in E. coli, the best-studied model bacterium. Comprehensive comparative studies of the targets revealed conserved target pathways of (p)ppGpp in both Gram-positive and -negative bacteria and novel targets of (p)ppGpp, including an alternative degradative pathway of (p)ppGpp. Thus, our discoveries may help in understanding of how (p)ppGpp increases the stress resilience and multidrug tolerance not only of the model organism E. coli but also of the pathogenic organisms in which these targets are conserved.

The sharp phase of respiratory inhibition during amino acid starvation in Escherichia coli is RelA-dependent and associated with regulation of ATP synthase activity

Amino acid starvation causes an RelA-dependent increase in the regulatory nucleotide (p)ppGpp that leads to pleiotropic changes in Escherichia coli metabolism, but the role of (p)ppGpp in regulation of respiration remains unclear. Here we demonstrate that amino acid starvation is accompanied by sharp RelA-dependent inhibition of respiration. The sharp phase of inhibition is absent in relA mutants, and can be prevented by translation inhibitors chloramphenicol and tetracycline, which abolish accumulation of (p)ppGpp. Single knockouts of any components of the respiratory chain do not affect inhibition of respiration. Studies of dO₂ changes in various atp mutants indicate that ATP synthase is probably the primary target of (p)ppGpp-mediated respiratory control. Inhibition of respiration induced by amino acid starvation is followed by transient perturbations in the membrane potential (Δψ) and K⁺ fluxes and leads to transient acceleration of superoxide production and H₂O₂ accumulation in the medium. High levels of H₂O₂ and superoxide formation and induced activity of antioxidant systems in the atpC mutant indicate the important role of ATP synthase in controlling the production of reactive oxygen species. The new function of (p)ppGpp, discovered here, expands the understanding of its role in metabolic reprogramming during the adaptive response to stresses.

Gene length as a regulator for ribosome recruitment and protein synthesis: theoretical insights

Protein synthesis rates are determined, at the translational level, by properties of the transcript’s sequence. The efficiency of an mRNA can be tuned by varying the ribosome binding sites controlling the recruitment of the ribosomes, or the codon usage establishing the speed of protein elongation. In this work we propose transcript length as a further key determinant of translation efficiency. Based on a physical model that considers the kinetics of ribosomes advancing on the mRNA and diffusing in its surrounding, as well as mRNA circularisation and ribosome drop-off, we explain how the transcript length may play a central role in establishing ribosome recruitment and the overall translation rate of an mRNA. According to our results, the proximity of the 3′ end to the ribosomal recruitment site of the mRNA could induce a feedback in the translation process that would favour the recycling of ribosomes. We also demonstrate how this process may be involved in shaping the experimental ribosome density-gene length dependence. Finally, we argue that cells could exploit this mechanism to adjust and balance the usage of its ribosomal resources.

Synchronized switching of multiple toxin-antitoxin modules by (p)ppGpp fluctuation

Toxin–antitoxin (TA) loci are widespread in bacteria including important pathogenic species. Recent studies suggest that TA systems play a key role in persister formation. However, the persistence phenotype shows only weak dependence on the number of TA systems, i.e. they are functionally redundant. We use a mathematical model to investigate the interaction of multiple TA systems in the switching between growth and persistence. We explore two scenarios: (i) TA systems are bistable and each TA system experiences its own noise and (ii) the noise in the level of common stress signal (e.g. (p)ppGpp) coordinates all TA systems simultaneously. We find that in the first scenario the exit from the persister state strongly depends on the number of TA systems. However in the second case, we could reproduce the weak dependence. The duration of the high (p)ppGpp state was found to be the key parameter for persistence. The (p)ppGpp-driven synchronized transition of all TA systems results in the redundancy.

Quantification of the Abundance and Charging Levels of Transfer RNAs in Escherichia coli

Transfer RNA (tRNA) is an essential part of the translational machinery in any organism. tRNAs bind and transfer amino acids to the translating ribosome. The relative levels of different tRNAs, and the ratio of aminoacylated tRNA to total tRNA, known as the charging level, are important factors in determining the accuracy and speed of translation. Therefore, the abundance and charging levels of tRNAs are important variables to measure when studying protein synthesis, for example under various stress conditions. Here, we describe a method for harvesting tRNA and directly measuring both the relative abundance and the absolute charging level of specific tRNA species in Escherichia coli. The tRNA is harvested in such a way that the labile bond between the tRNA and its amino acid is preserved. The RNA is then subjected to gel electrophoresis and Northern blotting, which results in separation of the charged and uncharged tRNAs. The levels of specific tRNAs in different samples can be compared due to the addition of spike-in cells for normalization. Prior to RNA purification, we add 5% of E. coli cells that overproduce the rare tRNA^selC to each sample. The amount of the tRNA species of interest in a sample is then normalized to the amount of tRNA^selC in the same sample. Addition of spike-in cells prior to RNA purification has the advantage over addition of purified spike-in RNAs that it also accounts for any differences in cell lysis efficiency between samples.

Mechanisms of bacterial persistence during stress and antibiotic exposure

Bacterial persister cells avoid antibiotic-induced death by entering a physiologically dormant state and are considered a major cause of antibiotic treatment failure and relapsing infections. Such dormant cells form stochastically, but also in response to environmental cues, by various pathways that are usually controlled by the second messenger (p)ppGpp. For example, toxin-antitoxin modules have been shown to play a major role in persister formation in many model systems. More generally, the diversity of molecular mechanisms driving persister formation is increasingly recognized as the cause of physiological heterogeneity that underlies collective multistress and multidrug tolerance of persister subpopulations. In this Review, we summarize the current state of the field and highlight recent findings, with a focus on the molecular basis of persister formation and heterogeneity.

Bacterial persistence from a system-level perspective

In recent years, our understanding about bacterial persistence has significantly advanced: we comprehend the persister phenotype better, more triggers for persistence entry have been found, and more insights in the involvement and role of toxin-antitoxin systems and other molecular mechanisms have been unravelled. In this review, we attempt to put these findings into an integrated, system-level perspective. From this point of view, persistence can be seen as a response to a strong perturbation of metabolic homeostasis, either triggered environmentally, or by means of intracellular stochasticity. Metabolic-flux-regulated resource allocation ensures stress protection, and several feedback mechanisms stabilize the cells in this protected state. We hope that this novel view can advance our understanding about persistence.

ATP-Dependent Persister Formation in Escherichia coli

Persisters are dormant variants that form a subpopulation of cells tolerant to antibiotics. Persisters are largely responsible for the recalcitrance of chronic infections to therapy. In Escherichia coli, one widely accepted model of persister formation holds that stochastic accumulation of ppGpp causes activation of the Lon protease that degrades antitoxins; active toxins then inhibit translation, resulting in dormant, drug-tolerant persisters. We found that various stresses induce toxin-antitoxin (TA) expression but that induction of TAs does not necessarily increase persisters. The 16S rRNA promoter rrnB P1 was proposed to be a persister reporter and an indicator of toxin activation regulated by ppGpp. Using fluorescence-activated cell sorting (FACS), we confirmed the enrichment for persisters in the fraction of rrnB P1-gfp dim cells; however, this is independent of toxin-antitoxins. rrnB P1 is coregulated by ppGpp and ATP. We show that rrnB P1 can report persisters in a relA/spoT deletion background, suggesting that rrnB P1 is a persister marker responding to ATP. Consistent with this finding, decreasing the level of ATP by arsenate treatment causes drug tolerance. Lowering ATP slows translation and prevents the formation of DNA double-strand breaks upon fluoroquinolone treatment. We conclude that variation in ATP levels leads to persister formation by decreasing the activity of antibiotic targets.

Persisters are a subpopulation of antibiotic-tolerant cells responsible for the recalcitrance of chronic infections. Our current understanding of persister formation is primarily based on studies of E. coli. The activation of toxin-antitoxin systems by ppGpp has become a widely accepted model for persister formation. In this study, we found that stress-induced activation of mRNA interferase-type toxins does not necessarily cause persister formation. We also found that the persister marker rrnB P1 reports persister cells because it detects a drop in cellular ATP levels. Consistent with this, lowering the ATP level decreases antibiotic target activity and, thus, leads to persister formation. We conclude that stochastic variation in ATP is the main mechanism of persister formation. A decrease in ATP provides a satisfactory explanation for the drug tolerance of persisters, since bactericidal antibiotics act by corrupting energy-dependent targets.

Transfer RNA is highly unstable during early amino acid starvation in Escherichia coli

Due to its long half-life compared to messenger RNA, bacterial transfer RNA is known as stable RNA. Here, we show that tRNAs become highly unstable as part of Escherichia coli's response to amino acid starvation. Degradation of the majority of cellular tRNA occurs within twenty minutes of the onset of starvation for each of several amino acids. Both the non-cognate and cognate tRNA for the amino acid that the cell is starving for are degraded, and both charged and uncharged tRNA species are affected. The alarmone ppGpp orchestrates the stringent response to amino acid starvation. However, tRNA degradation occurs in a ppGpp-independent manner, as it occurs with similar kinetics in a relaxed mutant. Further, we also observe rapid tRNA degradation in response to rifampicin treatment, which does not induce the stringent response. We propose a unifying model for these observations, in which the surplus tRNA is degraded whenever the demand for protein synthesis is reduced. Thus, the tRNA pool is a highly regulated, dynamic entity. We propose that degradation of surplus tRNA could function to reduce mistranslation in the stressed cell, because it would reduce competition between cognate and near-cognate charged tRNAs at the ribosomal A-site.