Bacterial Toxin–Antitoxin Systems: More Than Selfish Entities?

doi:10.1371/journal.pgen.1000437

PLoS Genet. 2009 March; 5(3): e1000437.

Published online 2009 March 27. doi: 10.1371/journal.pgen.1000437

PMCID: PMC2654758

Copyright Van Melderen, Saavedra De Bast. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Bacterial Toxin–Antitoxin Systems: More Than Selfish Entities?

Laurence Van Melderen^* and Manuel Saavedra De Bast

Laboratoire de Génétique et Physiologie Bactérienne, IBMM, Faculté des Sciences, Université Libre de Bruxelles, Gosselies, Belgium

Susan M. Rosenberg, Editor

Baylor College of Medicine, United States of America

* E-mail: lvmelder/at/ulb.ac.be

This article has been cited by other articles in PMC.

Abstract

Bacterial toxin–antitoxin (TA) systems are diverse and widespread in the prokaryotic kingdom. They are composed of closely linked genes encoding a stable toxin that can harm the host cell and its cognate labile antitoxin, which protects the host from the toxin's deleterious effect. TA systems are thought to invade bacterial genomes through horizontal gene transfer. Some TA systems might behave as selfish elements and favour their own maintenance at the expense of their host. As a consequence, they may contribute to the maintenance of plasmids or genomic islands, such as super-integrons, by post-segregational killing of the cell that loses these genes and so suffers the stable toxin's destructive effect. The function of the chromosomally encoded TA systems is less clear and still open to debate. This Review discusses current hypotheses regarding the biological roles of these evolutionarily successful small operons. We consider the various selective forces that could drive the maintenance of TA systems in bacterial genomes.

Introduction

Although bacteria have long been known to exchange genetic information through horizontal gene transfer, the impact of this dynamic process on genome evolution was fully appreciated only recently using comparative genomics (reviewed in [1]). Bacterial chromosomes are composed of genes that have quite different evolutionary origins (reviewed in [2]). The set of genes that is preferentially transmitted vertically over long evolutionary time scales composes the core genome. Core genes are relatively well conserved among different monophyletic groups and encode the cellular core functions. These core genes are interspersed with groups of genes that have been acquired from other prokaryotic genomes by horizontal transmission. These genomic islands mostly originate from integration events of mobile genetic elements, such as insertion sequences, transposons, phages, and plasmids. They might, therefore, be found in phylogenetically distant species and are not conserved among different isolates belonging to the same bacterial species. This set of genes constitutes the flexible genome.

Both gene influx and efflux processes are important in shaping bacterial-genome content. A vast majority of horizontally transferred genes are quickly lost after integration [3], although some remain interspersed in the genome (reviewed in [2]). Bacterial toxin-antitoxin (TA) systems appear to be subjected to this flux. Indeed, these small gene systems are found in plasmids as well as in chromosomes, and they are thought to be part of the flexible genome [4]. Although their role, when they are located in plasmid, is fairly clear, the involvement in physiological processes of the TA systems' chromosomally encoded counterparts is still open to debate.

Here we discuss current hypotheses regarding the biological roles of chromosomally encoded TA systems and consider the various selective forces that could drive the maintenance of TA systems in bacterial genomes.

Diversity and Abundance of Bacterial TA Systems

Bacterial TA systems are of two different types depending on the nature of the antitoxin; the toxin always being a protein. The antitoxin of type I systems is a small RNA (antisense or adjacent and divergent to the toxin gene) showing complementarity to the toxin mRNA (for recent reviews on type I systems, see [5],[6]). Type I antitoxins regulate toxin expression by inhibiting the toxin's translation. The toxins of type I systems are small, hydrophobic proteins that cause damage in bacterial cell membranes. In type II systems, the antitoxin is a small, unstable protein that sequesters the toxin through proteic complex formation (for a recent review on type II systems, see [7]). Much more information is available for type II systems, especially in terms of their biological roles. We will focus on the type II systems and use the term TA systems for brevity.

Type II TA systems are organised in operons, with the upstream gene usually encoding the antitoxin protein. The expression of the two genes is regulated at the level of transcription by the antitoxin–toxin complex. Nine families of toxins have been defined so far based on amino sequence homology [4]. Their targets and the cellular processes that are affected by their activities are shown in Table 1.

Table 1

The nine toxin families.

Comprehensive genome analyses have highlighted the diversity in the distribution of TA systems [4],[8],[9]. Some genomes such as that of Nitrosomonas europeae, Sinorhizobium meliloti, and Mycobacterium bovis contain more than 50 putative TA systems. Some others contain no or very few (less than three) putative TA systems, such as Rickettsia prowazeki, Campylobacter jejuni, or Bacillus subtilis. No correlation between the number of TA systems, the lifestyle, the membership of a phylum, or the growth rate (as it was proposed [4]) could be drawn [9]. Another level of diversity in distribution of TA systems among bacteria is added when comparing the occurrence of TA systems between different isolates of the same species. Table 2 shows the distribution of the nine toxins in seven sequenced Escherichia coli strains.

Table 2

Occurrence of toxin homologues in seven E. coli chromosomes.

As an example, homologues of the CcdB, MazF, and HipA toxins are frequently represented (at least in five chromosomes), whereas others appear to be absent (Doc and VapC) or present in only one chromosome (RelE and ParE). This implies that these TA systems were integrated in chromosomes through horizontal transfer, most probably in very recent events. The copy number of TA systems within one genome may also vary from one isolate to another. For instance, the MazE and HigB toxins are present in two copies in at least two genomes. Thus, TA systems are part of the flexible genome. They might be located in cryptic prophages such as relBE in the E. coli K-12 Qin prophage or constitute genomics islets by themselves such as ccd_O157 [10].

TA Systems: Just Selfish Entities?

Table 2 implies that the integrations of TA systems in E. coli chromosomes are recent events, because the distribution of the different TA systems varies from one isolate to another, raising the possibility that chromosomally encoded TA systems might have no physiological function. An attractive possibility is that TA systems act as selfish entities. Toxin and antitoxin genes show a strong interdependence (the functionality of the antitoxin is indispensable for the survival of cells carrying the toxin gene). They are closely linked, and they are capable of moving from one genome to another through horizontal gene transfer, as well as maintaining themselves in bacterial populations even at the expense of their host cell, at least when they are encoded in plasmids. Indeed, their stabilisation properties might be a consequence of their selfish behaviour (see below).

TA Systems: More than Selfish Entities?

Plasmid-Encoded TA Systems and Plasmid Fitness

Natural plasmids are often present in bacteria at very low copy number (one copy per chromosome). They are also able to spread by conjugation or mobilization with the help of other conjugative plasmids. They thus constitute a substantial proportion of the flexible genome and contribute importantly to bacterial evolution. TA systems increase the plasmid prevalence (number of plasmid-containing cells/total number of cells) in growing bacterial populations by selectively eliminating daughter cells that did not inherit a plasmid copy at cell division [11],[12] (Figure 1A). This post-segregational killing mechanism relies on the differential stability of the toxin and antitoxin [13],[14]. In daughter bacteria devoid of a plasmid copy, because TA proteins are not replenished, the antitoxin pool rapidly decreases, freeing the stable toxin. These plasmid-free bacteria will eventually be killed by the deleterious activity of the toxin. Plasmid-encoded TA systems are also called addiction modules [15], since this property renders the cell addicted to antitoxin production and therefore to the TA genes. Cooper and Heinemann [16] showed that TA systems might also function in plasmid–plasmid competition, as proposed for restriction-modification systems in the “selfish theory” of Kobayashi and colleagues [17],[18]. They showed that plasmid-encoded TA systems allow a conjugative plasmid (PSK⁺ plasmid) to outcompete a conjugative plasmid belonging to the same incompatibility group (identical replicon) but devoid of the TA system (PSK⁻ plasmid) [16]. Therefore, TA systems increase the relative fitness of their host DNA molecules by eliminating competitor plasmids in the bacterial progeny through post-segregational killing (Figure 1B). Mathematical models demonstrate that the post-segregational killing phenomenon allows the propagation of TA systems in bacterial populations, independently of their original frequencies [19]. This might provide rational explanation for the evolutionary success of TA systems.

Figure 1

Advantage conferred by plasmid-encoded TA systems.

Chromosomally Encoded TA Systems

Some chromosomally encoded TA systems might be integrated in host regulatory networks and thereby confer a fitness advantage to the bacterial-host cells and/or populations. Several models supporting this view have been proposed.

A secure way to survive: Being integrated in host regulatory networks?

The programmed-cell death model is based on the study of the chromosomally encoded mazEF TA system of E. coli (reviewed in [20]). mazEF-mediated programmed cell death was observed by Engelberg-Kulka and colleagues under a wide variety of unrelated stressful conditions (e.g., amino-acid starvation, short-term antibiotic treatments, high temperature, and oxidative shock). Stress conditions are thought to affect the production of the mazEF-encoded proteins in a manner dependent on ppGpp, an alarmone synthesised under starvation [21]–[24] and through a quorum-sensing-like small peptide (extra-cellular death factor or EDF) [25]. This particular combination of stress conditions and EDF is thought to shut off mazEF transcription and lead to MazF toxin liberation as a consequence of MazE degradation by the ClpAP ATP-dependent protease. The outcome of this activation has been shown to be fatal for at least 95% of the bacterial population. Altruistic death of a fraction of the bacterial population is proposed to provide nutriments for the siblings. The molecular mechanisms underlying this proposed stochastic activation, as well as those by which killing is achieved, are still unknown. Whether MazF induces cell lysis also remains to be established.

The growth-modulation model is built on data mostly obtained on the E. coli relBE system and to a lesser extent on mazEF and chpB (which encodes a toxin homologous to MazF) [26],[27]. This model relies on the primary observation that amino-acid starvation inhibits cell growth without leading to cell death [26], in contrast with the programmed cell-death model. However, growth inhibition was subsequently shown to be independent of the presence of relBE, mazEF, chpB, and two other type II systems [28]. Nevertheless, upon amino-acid starvation, the rate of translation drastically drops in a wild-type E. coli strain and to a lesser extent in a ΔrelBE mutant strain [26]. Gerdes and collaborators therefore proposed that relBE is a stress-response module that functions in quality control of gene expression to regulate the global level of translation, together with the trans-translation ssrA system [29]. Amino acid starvation activates relBE transcription through the Lon-dependent degradation of RelB and in a ppGpp-independent manner. As a consequence, RelE inhibits translation and induces a dormant state until favourable growth conditions return. Data obtained on mazEF and chpB by the group of Gerdes are consistent with the growth-regulator model and disagree with the programmed cell-death model [27], although each model could be true under different circumstances [21].

The persistence model describes an epigenetic trait that allows a small fraction of bacteria (~10⁻⁶) to enter into a dormant state that renders them able to survive stress conditions, notably antibiotic treatments (reviewed in [30]). A nontoxic mutant of the HipA toxin (hipA7) has been shown to confer high persistence in E. coli [31]. Mutations abolishing the production of the ppGpp alarmone eliminated the high persistence phenotype, suggesting that hipA7 might induce a high level of ppGpp [31]. Persistence and toxicity might be independent, because the HipA7 mutant seems to be less efficient for inhibition of macromolecule synthesis as compared to the wild-type HipA [32]. However, the protein kinase activity of HipA was shown to be required for persistence and growth arrest [33]. The central elongation factor Tu (EF-Tu) was recently shown to bind and to be phosphorylated by HipA [34]. EF-Tu in its nonphospohorylated form catalyses the binding of aminoacyl-tRNAs to the ribosome. Phosphorylation of EF-Tu by HipA might lead to translation inhibition [34] and therefore to ppGpp synthesis. Single-cell analysis revealed that several TA systems are up-regulated in persister cells [35]. The biological meaning of this observation remains unclear, since the deletion mazEF and relBE did not impair persister frequency under ofloxacin (a fluoroquinolone) or mitomycin C treatments. However, the ΔhipBA mutant strain was strongly affected (10- to 100-fold), showing that this TA system is involved in persistence [36]. The molecular mechanisms underlying this stochastic phenomenon are unknown.

The development model was proposed recently for fruiting body formation in Myxococcus xanthus. A homologue of the mazF toxin gene (mazF-mx), which is devoid of any mazE antitoxin gene homologue, was identified in the chromosome of M. xanthus [37]. The solitary mazF-mx toxin gene constitutes an interesting example of integration in host regulatory networks. M. xanthus forms multicellular structures called fruiting bodies under nutrient-starvation conditions. During this process, 80% of the population engaged in fruiting-body formation die by lysis; only 20% will develop into myxospores. The mazF-mx gene is integrated in a regulatory cascade controlled by the key developmental regulator MrpC, which presents a dual activity towards mazF-mx: it positively regulates mazF-mx expression at the transcriptional level and it negatively controls its endoribonuclease activity at the post-translational level by acting as its antitoxin. During vegetative growth, MrpC transcriptional activity is controlled negatively by its phosphorylation through a Ser/Thr protein kinase. When M. xanthus engages in fruiting body formation, MrpC transcription activity is activated most likely by a LonD-dependent cleavage. MazF-mx is then produced and cleaves mRNAs, thereby inducing cell death. mazF-mx is essential for fruiting body formation, because a ΔmazF-mx mutant shows a dramatic reduction of myxospore formation.

In the above models, chromosomally encoded TA systems are thought to be integral parts of their host genetic networks. mazEF has been extensively reported as being responsible for programmed cell death, although this observation failed to be reproduced in various labs and is still a subject of debate [26]–[28]. Nevertheless, TA systems are thought to allow cells and/or populations to cope with stress conditions, and should therefore confer a clear selective advantage in these conditions. Indeed, mazF-mx and hipBA appear to be essential components of host regulatory networks, since their deletion caused a drastic phenotype [36],[37]. However, it is less clear for mazEF, relBE, and chpB of E. coli, since no fitness gain could be attributed to their presence neither under stress conditions nor during post-stress recovery phases [28].

The two following models provide an alternative to the previous ones by illustrating how TA systems can confer selective advantages to their bacterial host without being integrated into regulatory networks.

TA systems in dynamic genome evolution

The stabilisation model proposes that because of their addictive characteristics, chromosomally encoded TA systems could act against large-scale deletion of otherwise dispensable genomic regions [38]. Super-integrons are plastic platforms composed of numerous gene cassettes (more than a hundred in the Vibrio cholerae super-integron) and repeat sequences (reviewed in [39]). Super-integrons encode many functions (e.g., antibiotic resistance). Super-integrons may advantage bacterial populations over long time scales by maintaining nonessential genes and allowing bacterial lineages to better cope with unpredictable changes of environmental conditions. Gene cassettes are excised, integrated, and rearranged by the action of the SI-encoded integrase. They contain in general a single gene devoid of promoter, except for the TA systems encoding cassettes. In this case, the entire TA operon is present in the cassette and is most likely expressed. Several TA systems from super-integrons belonging to various Vibrionaceae are able to stabilise otherwise unstable plasmids or large genomic regions in E. coli [38],[40],[41]. Moreover, super-integrons are extremely stable. Attempts to delete the super-integron of V. cholerae have failed, strongly suggesting that TA systems serve to stabilise the super-integron platform and counteract gene efflux (D. Mazel, personal communication).

While it becomes clear that TA systems in such genetic structures or in cryptic prophages such as relBE of Qin [42] have retained their stabilisation properties, the generalisation to more “classical” chromosomally encoded TA systems should be taken with caution. Although only a few systems have been tested (E. coli dinJ-yafQ and ccd_O157 systems), they appear to be unable to prevent large-scale deletion or to stabilise an otherwise unstable plasmid [10],[38]. Wide surveys of stabilisation properties of TA systems from various locations (mobile genetic elements, core, genomic islands, remnants) will test whether a correlation between stabilisation function and localisation exists.

The anti-addiction model proposes that chromosomally-encoded systems can selectively advantage their host in post-segregational killing conditions. In theory, chromosomally-encoded antitoxins sharing sufficient identity with homologous plasmid-encoded TA systems might act as anti-addiction modules by preventing post-segregational killing (Figure 2). The ccd_Ech chromosomally encoded TA system of Erwinia chrysanthemi 3937 was shown to have this property with respect to its E. coli F plasmid–encoded ccd_F homolog [43]. In an E. coli strain containing the ccd_Ech system inserted in its chromosome (ccd_Ech strain), no post-segregational killing was observed upon the loss of a plasmid carrying ccd_F. Moreover, competition experiments showed that under post-segregational killing conditions, the ccd_Ech strain had a selective advantage compared to the wild-type strain. Therefore, the fitness advantage conferred by the newly acquired anti-addiction module under post-segregational killing conditions might allow its fixation in the bacterial population. In turn, the plasmid-encoded system will lose its addictive character. On the one hand, variants able to evade anti-addiction modules are expected to be selected and out-compete their post-segregational killing–defective relatives. Anti-addiction might thus be one of the evolutionary forces driving selection of the plasmid encoded TA systems. On the other hand, chromosomally encoded TA systems might lose their anti-addictive properties [10] and decay [44].

Figure 2

The anti-addiction model.

Conclusions

There is no doubt that bacterial TA systems are evolutionarily successful entities. Some bacterial genomes harbour several dozen of them [4],[9]. Even obligatorily intracellular species that undergo massive genome reduction contain TA systems [9],[45]. There is increasing evidence that these small entities move between genomes through horizontal gene transfer. Their phylogeny is not congruent with the bacterial one [4],[46], and their distribution varies greatly between isolates belonging to the same bacterial species ([44],[46], Table 2), implying that TA systems are highly mobile. Pandey and Gerdes also reported recently that TA systems are preferentially associated with genomic islands [4]. However, how horizontally acquired TA systems are fixed within the population is not yet understood. One can argue that their addictive “selfish” characteristics enable them to be stabilised and refractory to gene efflux. As a consequence, in specific genomic locations such as plasmids or genomic islands, they may contribute to the maintenance of these structures in bacterial population by post-segregational killing and be subjected to selection. In other genomic locations, such as the core genome where they are not subjected to selection, some TA systems might accumulate mutations that reduce or inactivate their addictive properties simply by genetic drift. Indeed, deletion of both type II [24],[26],[47] and type I systems [48],[49] in E. coli K-12 was possible, at least under the conditions used in these experiments, suggesting that these systems have lost their addictive characteristics. Signs of “loss of addictive properties” were detected for several type I and II systems. For instance, the five copies of the type I hok-sok system located in the E. coli K-12 chromosome are inactivated by insertion sequences, point mutation, or large rearrangements [50], and the ccd_O157 system appears to undergo a degenerative process within the E. coli species [44]. Similar observations have been reported for restriction-modification systems that share the addiction and apparent mobility characteristics of TA systems [51],[52]. Another route for TA system evolution is their integration into host regulatory networks. This is exemplified by the MazF-mx toxin in M. xanthus that had been hijacked by the developmental network controlling fruiting-body formation. The canonical antitoxin has been replaced by a complex cascade of signal transduction proteins involving a Ser/Thr protein kinase and a transcriptional activator/antitoxin protein [37].

Many scenarios might occur depending notably on the bacterial species and the type of toxin. The TA field should avoid generalisation regarding the biological role of these interesting entities. These small modules are highly diverse and ubiquitous. They might have multiple biological roles, if any, that depend on their age, their genomic location, the nature of the toxin, and most likely on many not-yet-discovered factors that influence their evolution.

Acknowledgments

We thank the past and present members of the laboratory for constant discussions.

Footnotes

The authors have declared that no competing interests exist.

The authors received no specific funding for this article.

References

1. Doolittle RF. Evolutionary aspects of whole-genome biology. Curr Opin Struct Biol. 2005;15:248–253. [PubMed]

2. Lawrence JG, Hendrickson H. Genome evolution in bacteria: order beneath chaos. Curr Opin Microbiol. 2005;8:572–578. [PubMed]

3. Kunin V, Ouzounis CA. The balance of driving forces during genome evolution in prokaryotes. Genome Res. 2003;13:1589–1594. [PubMed]

4. Pandey DP, Gerdes K. Toxin-antitoxin loci are highly abundant in free-living but lost from host-associated prokaryotes. Nucleic Acids Res. 2005;33:966–976. [PMC free article] [PubMed]

5. Fozo EM, Hemm MR, Storz G. Small toxic proteins and the antisense RNAs that repress them. Microbiol Mol Biol Rev. 2008;72:579–589. [PMC free article] [PubMed]

6. Gerdes K, Wagner EG. RNA antitoxins. Curr Opin Microbiol. 2007;10:117–124. [PubMed]

7. Gerdes K, Christensen SK, Lobner-Olesen A. Prokaryotic toxin-antitoxin stress response loci. Nat Rev Microbiol. 2005;3:371–382. [PubMed]

8. Anantharaman V, Aravind L. New connections in the prokaryotic toxin-antitoxin network: relationship with the eukaryotic nonsense-mediated RNA decay system. Genome Biol. 2003;4:R81. [PMC free article] [PubMed]

9. Sevin EW, Barloy-Hubler F. RASTA-Bacteria: a web-based tool for identifying toxin-antitoxin loci in prokaryotes. Genome Biol. 2007;8:R155. [PMC free article] [PubMed]

10. Wilbaux M, Mine N, Guerout AM, Mazel D, Van Melderen L. Functional interactions between coexisting toxin-antitoxin systems of the ccd family in Escherichia coli O157:H7. J Bacteriol. 2007;189:2712–2719. [PMC free article] [PubMed]

11. Gerdes K, Rasmussen PB, Molin S. Unique type of plasmid maintenance function: postsegregational killing of plasmid-free cells. Proc Natl Acad Sci U S A. 1986;83:3116–3120. [PubMed]

12. Jaffe A, Ogura T, Hiraga S. Effects of the ccd function of the F plasmid on bacterial growth. J Bacteriol. 1985;163:841–849. [PMC free article] [PubMed]

13. Tsuchimoto S, Nishimura Y, Ohtsubo E. The stable maintenance system pem of plasmid R100: degradation of PemI protein may allow PemK protein to inhibit cell growth. J Bacteriol. 1992;174:4205–4211. [PMC free article] [PubMed]

14. Van Melderen L, Bernard P, Couturier M. Lon-dependent proteolysis of CcdA is the key control for activation of CcdB in plasmid-free segregant bacteria. Mol Microbiol. 1994;11:1151–1157. [PubMed]

15. Yarmolinsky MB. Programmed cell death in bacterial populations. Science. 1995;267:836–837. [PubMed]

16. Cooper TF, Heinemann JA. Postsegregational killing does not increase plasmid stability but acts to mediate the exclusion of competing plasmids. Proc Natl Acad Sci U S A. 2000;97:12643–12648. [PubMed]

17. Naito T, Kusano K, Kobayashi I. Selfish behavior of restriction-modification systems. Science. 1995;267:897–899. [PubMed]

18. Naito Y, Naito T, Kobayashi I. Selfish restriction modification genes: resistance of a resident R/M plasmid to displacement by an incompatible plasmid mediated by host killing. Biol Chem. 1998;379:429–436. [PubMed]

19. Mochizuki A, Yahara K, Kobayashi I, Iwasa Y. Genetic addiction: selfish gene's strategy for symbiosis in the genome. Genetics. 2006;172:1309–1323. [PubMed]

20. Engelberg-Kulka H, Amitai S, Kolodkin-Gal I, Hazan R. Bacterial programmed cell death and multicellular behavior in bacteria. PLoS Genet. 2006;2:e135. doi:10.1371/journal.pgen.0020135. [PMC free article] [PubMed]

21. Amitai S, Yassin Y, Engelberg-Kulka H. MazF-mediated cell death in Escherichia coli: a point of no return. J Bacteriol. 2004;186:8295–8300. [PMC free article] [PubMed]

22. Hazan R, Sat B, Engelberg-Kulka H. Escherichia coli mazEF-mediated cell death is triggered by various stressful conditions. J Bacteriol. 2004;186:3663–3669. [PMC free article] [PubMed]

23. Sat B, Hazan R, Fisher T, Khaner H, Glaser G, et al. Programmed cell death in Escherichia coli: some antibiotics can trigger mazEF lethality. J Bacteriol. 2001;183:2041–2045. [PMC free article] [PubMed]

24. Aizenman E, Engelberg-Kulka H, Glaser G. An Escherichia coli chromosomal “addiction module” regulated by guanosine [corrected] 3′,5′-bispyrophosphate: a model for programmed bacterial cell death. Proc Natl Acad Sci U S A. 1996;93:6059–6063. [PubMed]

25. Kolodkin-Gal I, Hazan R, Gaathon A, Carmeli S, Engelberg-Kulka H. A linear pentapeptide is a quorum-sensing factor required for mazEF-mediated cell death in Escherichia coli. Science. 2007;318:652–655. [PubMed]

26. Christensen SK, Mikkelsen M, Pedersen K, Gerdes K. RelE, a global inhibitor of translation, is activated during nutritional stress. Proc Natl Acad Sci U S A. 2001;98:14328–14333. [PubMed]

27. Christensen SK, Pedersen K, Hansen FG, Gerdes K. Toxin-antitoxin loci as stress-response-elements: ChpAK/MazF and ChpBK cleave translated RNAs and are counteracted by tmRNA. J Mol Biol. 2003;332:809–819. [PubMed]

28. Tsilibaris V, Maenhaut-Michel G, Mine N, Van Melderen L. What is the benefit to Escherichia coli of having multiple toxin-antitoxin systems in its genome? J Bacteriol. 2007;189:6101–6108. [PMC free article] [PubMed]

29. Christensen SK, Gerdes K. RelE toxins from bacteria and Archaea cleave mRNAs on translating ribosomes, which are rescued by tmRNA. Mol Microbiol. 2003;48:1389–1400. [PubMed]

30. Lewis K. Persister cells and the riddle of biofilm survival. Biochemistry (Mosc) 2005;70:267–274. [PubMed]

31. Korch SB, Henderson TA, Hill TM. Characterization of the hipA7 allele of Escherichia coli and evidence that high persistence is governed by (p)ppGpp synthesis. Mol Microbiol. 2003;50:1199–1213. [PubMed]

32. Korch SB, Hill TM. Ectopic overexpression of wild-type and mutant hipA genes in Escherichia coli: effects on macromolecular synthesis and persister formation. J Bacteriol. 2006;188:3826–3836. [PMC free article] [PubMed]

33. Correia FF, D'Onofrio A, Rejtar T, Li L, Karger BL, et al. Kinase activity of overexpressed HipA is required for growth arrest and multidrug tolerance in Escherichia coli. J Bacteriol. 2006;188:8360–8367. [PMC free article] [PubMed]

34. Schumacher MA, Piro KM, Xu W, Hansen S, Lewis K, et al. Molecular mechanisms of HipA-mediated multidrug tolerance and its neutralization by HipB. Science. 2009;323:396–401. [PMC free article] [PubMed]

35. Shah D, Zhang Z, Khodursky A, Kaldalu N, Kurg K, et al. Persisters: a distinct physiological state of E. coli. BMC Microbiol. 2006;6:53. [PMC free article] [PubMed]

36. Keren I, Shah D, Spoering A, Kaldalu N, Lewis K. Specialized persister cells and the mechanism of multidrug tolerance in Escherichia coli. J Bacteriol. 2004;186:8172–8180. [PMC free article] [PubMed]

37. Nariya H, Inouye M. MazF, an mRNA interferase, mediates programmed cell death during multicellular Myxococcus development. Cell. 2008;132:55–66. [PubMed]

38. Rowe-Magnus DA, Guerout AM, Biskri L, Bouige P, Mazel D. Comparative analysis of superintegrons: engineering extensive genetic diversity in the Vibrionaceae. Genome Res. 2003;13:428–442. [PubMed]

39. Mazel D. Integrons: agents of bacterial evolution. Nat Rev Microbiol. 2006;4:608–620. [PubMed]

40. Christensen-Dalsgaard M, Gerdes K. Two higBA loci in the Vibrio cholerae superintegron encode mRNA cleaving enzymes and can stabilize plasmids. Mol Microbiol. 2006;62:397–411. [PubMed]

41. Szekeres S, Dauti M, Wilde C, Mazel D, Rowe-Magnus DA. Chromosomal toxin-antitoxin loci can diminish large-scale genome reductions in the absence of selection. Mol Microbiol. 2007;63:1588–1605. [PubMed]

42. Gotfredsen M, Gerdes K. The Escherichia coli relBE genes belong to a new toxin-antitoxin gene family. Mol Microbiol. 1998;29:1065–1076. [PubMed]

43. Saavedra De Bast M, Mine N, Van Melderen L. Chromosomal toxin-antitoxin systems may act as antiaddiction modules. J Bacteriol. 2008;190:4603–4609. [PMC free article] [PubMed]

44. Mine N, Guglielmini J, Wilbaux M, Van Melderen L. The decay of the chromosomally-encoded ccdO157 toxin-antitoxin system in the E. coli species. Genetics. 2009 E-pub ahead of print. doi:10.1534/genetics.108.095190. [PubMed]

45. Ogata H, Renesto P, Audic S, Robert C, Blanc G, et al. The genome sequence of Rickettsia felis identifies the first putative conjugative plasmid in an obligate intracellular parasite. PLoS Biol. 2005;3:e248. doi:10.1371/journal.pbio.0030248. [PMC free article] [PubMed]

46. Magnuson RD. Hypothetical functions of toxin-antitoxin systems. J Bacteriol. 2007;189:6089–6092. [PMC free article] [PubMed]

47. Christensen SK, Maenhaut-Michel G, Mine N, Gottesman S, Gerdes K, et al. Overproduction of the Lon protease triggers inhibition of translation in Escherichia coli: involvement of the yefM-yoeB toxin-antitoxin system. Mol Microbiol. 2004;51:1705–1717. [PubMed]

48. Fozo EM, Kawano M, Fontaine F, Kaya Y, Mendieta KS, et al. Repression of small toxic protein synthesis by the Sib and OhsC small RNAs. Mol Microbiol. 2008;70:1076–1093. [PMC free article] [PubMed]

49. Kawano M, Oshima T, Kasai H, Mori H. Molecular characterization of long direct repeat (LDR) sequences expressing a stable mRNA encoding for a 35-amino-acid cell-killing peptide and a cis-encoded small antisense RNA in Escherichia coli. Mol Microbiol. 2002;45:333–349. [PubMed]

50. Pedersen K, Gerdes K. Multiple hok genes on the chromosome of Escherichia coli. Mol Microbiol. 1999;32:1090–1102. [PubMed]

51. Kong H, Lin LF, Porter N, Stickel S, Byrd D, et al. Functional analysis of putative restriction-modification system genes in the Helicobacter pylori J99 genome. Nucleic Acids Res. 2000;28:3216–3223. [PMC free article] [PubMed]

52. Nobusato A, Uchiyama I, Ohashi S, Kobayashi I. Insertion with long target duplication: a mechanism for gene mobility suggested from comparison of two related bacterial genomes. Gene. 2000;259:99–108. [PubMed]

53. Garcia-Pino A, Christensen-Dalsgaard M, Wyns L, Yarmolinsky M, Magnuson RD, et al. Doc of Prophage P1 Is Inhibited by Its Antitoxin Partner Phd through Fold Complementation. J Biol Chem. 2008;283:30821–30827. [PubMed]

54. Liu M, Zhang Y, Inouye M, Woychik NA. Bacterial addiction module toxin Doc inhibits translation elongation through its association with the 30S ribosomal subunit. Proc Natl Acad Sci U S A. 2008;105:5885–5890. [PubMed]

55. Daines DA, Wu MH, Yuan SY. VapC-1 of nontypeable Haemophilus influenzae is a ribonuclease. J Bacteriol. 2007;189:5041–5048. [PMC free article] [PubMed]

56. Meinhart A, Alings C, Strater N, Camacho AG, Alonso JC, et al. Crystallization and preliminary X-ray diffraction studies of the epsilonzeta addiction system encoded by Streptococcus pyogenes plasmid pSM19035. Acta Crystallogr D Biol Crystallogr. 2001;57:745–747. [PubMed]

57. Lioy VS, Martin MT, Camacho AG, Lurz R, Antelmann H, et al. pSM19035-encoded zeta toxin induces stasis followed by death in a subpopulation of cells. Microbiology. 2006;152:2365–2379. [PubMed]

58. Budde PP, Davis BM, Yuan J, Waldor MK. Characterization of a higBA toxin-antitoxin locus in Vibrio cholerae. J Bacteriol. 2007;189:491–500. [PMC free article] [PubMed]

59. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, et al. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997;25:3389–3402. [PMC free article] [PubMed]

Articles from PLoS Genetics are provided here courtesy of
Public Library of Science